diff --git "a/model.safetensors.index.json" "b/model.safetensors.index.json"
new file mode 100644--- /dev/null
+++ "b/model.safetensors.index.json"
@@ -0,0 +1,11691 @@
+{
+  "metadata": {
+    "total_size": 26286253792
+  },
+  "weight_map": {
+    "language_model.lm_head.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.embed_tokens.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.0.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.1.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.10.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.11.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.12.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.13.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.14.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.15.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.16.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.17.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.18.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.19.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.2.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.20.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.21.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.22.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.23.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.24.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.25.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.26.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.27.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.28.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.29.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.3.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.30.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.31.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.32.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.33.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.34.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.35.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.36.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.37.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.38.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.39.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.4.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.40.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.41.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.42.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.43.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.44.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.45.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.46.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.47.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.48.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.49.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.5.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.50.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.51.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.52.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.53.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.54.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.55.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.56.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.57.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.58.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.59.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.6.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.60.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.61.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.62.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.63.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.7.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.8.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.experts.experts.9.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.router.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.mlp.shared_experts.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.k_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.o_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.q_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.0.self_attn.v_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.0.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.1.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.10.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.11.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.12.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.13.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.14.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.15.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.16.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.17.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.18.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.19.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.2.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.20.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.21.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.22.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.23.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.24.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.25.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.26.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.27.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.28.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.29.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.3.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.30.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.31.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.32.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.33.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.34.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.35.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.36.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.37.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.38.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.39.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.4.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.40.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.41.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.42.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.43.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.44.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.45.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.46.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.47.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.48.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.49.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.5.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.50.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.51.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.52.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.53.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.54.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.55.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.56.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.57.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.58.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.59.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.6.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.60.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.61.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.62.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.63.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.7.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.8.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.experts.experts.9.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.router.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.mlp.shared_experts.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.k_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.o_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.q_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.1.self_attn.v_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.10.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.40.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.41.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.42.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.43.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.44.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.45.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.46.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.47.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.48.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.49.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.50.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.51.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.52.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.53.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.54.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.55.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.56.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.57.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.10.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.40.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.41.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.42.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.43.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.44.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.45.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.46.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.47.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.48.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.49.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.50.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.51.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.52.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.53.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.54.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.55.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.56.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.57.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.11.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.40.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.41.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.42.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.43.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.44.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.45.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.46.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.47.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.48.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.49.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.50.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.51.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.52.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.53.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.54.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.55.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.56.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.57.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.12.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.40.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.41.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.42.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.43.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.44.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.45.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.46.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.47.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.48.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.49.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.50.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.51.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.52.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.53.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.54.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.55.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.56.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.57.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.13.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.40.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.41.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.42.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.43.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.44.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.45.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.46.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.47.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.48.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.49.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.50.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.51.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.52.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.53.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.54.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.55.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.56.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.57.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.14.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.0.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.1.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.10.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.11.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.12.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.13.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.14.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.15.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.16.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.17.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.18.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.19.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.2.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.20.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.21.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.22.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.23.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.24.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.25.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.26.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.27.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.28.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.29.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.3.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.30.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.31.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.32.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.33.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.34.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.35.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.36.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.37.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.38.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.39.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.4.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.5.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.6.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.7.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.8.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.experts.experts.9.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.router.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.k_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.k_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.o_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.o_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.q_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.q_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.v_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.15.self_attn.v_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.16.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.23.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.24.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.25.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.26.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.27.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.28.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.29.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.30.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.31.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.32.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.33.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.34.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.35.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.36.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.37.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.38.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.39.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.16.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.23.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.24.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.25.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.26.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.27.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.28.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.29.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.30.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.31.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.32.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.33.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.34.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.35.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.36.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.37.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.38.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.39.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.17.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.23.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.24.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.25.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.26.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.27.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.28.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.29.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.30.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.31.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.32.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.33.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.34.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.35.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.36.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.37.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.38.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.39.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.18.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.23.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.24.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.25.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.26.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.27.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.28.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.29.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.30.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.31.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.32.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.33.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.34.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.35.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.36.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.37.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.38.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.39.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.19.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.2.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.0.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.1.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.10.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.11.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.12.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.13.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.14.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.15.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.16.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.17.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.18.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.19.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.2.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.20.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.21.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.22.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.23.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.24.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.25.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.26.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.27.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.28.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.29.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.3.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.30.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.31.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.32.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.33.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.34.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.35.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.36.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.37.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.38.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.39.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.4.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.40.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.41.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.42.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.43.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.44.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.45.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.46.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.47.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.48.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.49.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.5.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.50.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.51.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.52.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.53.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.54.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.55.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.56.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.57.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.58.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.59.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.6.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.60.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.61.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.62.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.63.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.7.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.8.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.experts.experts.9.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.router.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.mlp.shared_experts.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.k_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.o_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.q_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.2.self_attn.v_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.20.input_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.23.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.24.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.25.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.26.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.27.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.28.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.29.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.30.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.31.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.32.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.33.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.34.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.35.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.36.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.37.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.38.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.39.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.40.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.41.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.42.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.43.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.44.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.45.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.46.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.47.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.48.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.49.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.50.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.51.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.52.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.53.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.54.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.55.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.56.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.57.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.58.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.59.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.60.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.61.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.62.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.63.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.mlp.shared_experts.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.post_attention_layernorm.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.20.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.0.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.1.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.10.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.11.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.12.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.13.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.14.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.15.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.16.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.17.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.18.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.19.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.2.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.20.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.21.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.22.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.3.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.4.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.5.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.6.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.7.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.8.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.down_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.down_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.gate_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.gate_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.up_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.experts.experts.9.up_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.router.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.k_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.k_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.o_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.o_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.q_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.q_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.v_proj.weight": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.21.self_attn.v_proj.weight_scale": "model-00004-of-00006.safetensors",
+    "language_model.model.layers.22.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.10.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.11.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.12.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.13.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.14.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.15.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.16.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.17.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.18.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.19.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.20.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.21.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.22.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.5.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.6.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.7.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.8.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.experts.experts.9.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.22.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.10.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.11.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.12.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.13.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.14.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.15.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.16.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.17.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.18.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.19.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.20.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.21.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.22.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.5.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.6.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.7.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.8.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.experts.experts.9.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.23.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.10.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.11.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.12.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.13.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.14.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.15.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.16.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.17.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.18.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.19.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.20.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.21.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.22.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.5.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.6.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.7.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.8.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.experts.experts.9.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.24.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.10.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.11.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.12.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.13.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.14.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.15.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.16.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.17.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.18.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.19.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.20.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.21.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.22.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.5.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.6.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.7.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.8.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.experts.experts.9.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.25.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.input_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.10.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.11.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.12.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.13.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.14.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.15.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.16.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.17.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.18.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.19.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.20.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.21.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.22.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.23.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.24.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.25.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.26.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.27.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.28.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.29.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.30.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.31.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.32.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.33.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.34.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.35.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.36.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.37.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.38.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.39.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.40.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.41.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.42.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.43.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.44.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.45.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.46.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.47.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.48.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.49.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.5.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.50.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.51.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.52.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.53.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.54.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.55.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.56.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.57.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.58.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.59.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.6.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.60.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.61.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.62.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.63.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.7.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.8.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.experts.experts.9.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.mlp.shared_experts.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.post_attention_layernorm.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.26.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.input_layernorm.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.0.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.1.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.10.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.11.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.12.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.13.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.14.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.15.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.16.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.17.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.18.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.19.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.2.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.20.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.21.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.22.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.23.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.24.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.25.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.26.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.27.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.28.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.29.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.3.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.30.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.31.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.32.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.33.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.34.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.35.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.36.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.37.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.38.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.39.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.down_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.down_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.gate_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.gate_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.up_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.4.up_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.40.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.41.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.42.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.43.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.44.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.45.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.46.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.47.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.48.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.49.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.5.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.50.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.51.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.52.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.53.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.54.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.55.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.56.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.57.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.58.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.59.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.6.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.60.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.61.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.62.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.63.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.7.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.8.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.experts.experts.9.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.router.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.down_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.down_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.gate_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.gate_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.up_proj.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.mlp.shared_experts.up_proj.weight_scale": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.post_attention_layernorm.weight": "model-00006-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.k_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.k_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.o_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.o_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.q_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.q_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.v_proj.weight": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.27.self_attn.v_proj.weight_scale": "model-00005-of-00006.safetensors",
+    "language_model.model.layers.3.input_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.0.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.1.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.10.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.11.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.12.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.13.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.14.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.15.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.16.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.17.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.18.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.19.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.2.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.20.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.21.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.22.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.23.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.24.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.25.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.26.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.27.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.28.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.29.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.3.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.30.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.31.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.32.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.33.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.34.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.35.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.36.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.37.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.38.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.39.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.4.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.40.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.41.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.42.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.43.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.44.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.45.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.46.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.47.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.48.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.49.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.5.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.50.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.51.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.52.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.53.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.54.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.55.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.56.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.57.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.58.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.59.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.6.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.60.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.61.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.62.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.63.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.7.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.8.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.experts.experts.9.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.router.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.mlp.shared_experts.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.k_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.o_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.q_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.3.self_attn.v_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.0.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.1.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.2.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.3.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.4.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.5.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.58.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.59.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.down_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.down_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.6.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.60.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.61.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.62.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.63.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.gate_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.gate_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.up_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.7.up_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.router.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.mlp.shared_experts.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.k_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.o_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.q_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.4.self_attn.v_proj.weight_scale": "model-00001-of-00006.safetensors",
+    "language_model.model.layers.5.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.0.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.1.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.2.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.3.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.4.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.5.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.58.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.59.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.6.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.60.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.61.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.62.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.63.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.7.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.router.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.mlp.shared_experts.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.5.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.0.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.1.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.2.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.3.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.4.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.5.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.58.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.59.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.6.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.60.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.61.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.62.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.63.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.7.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.router.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.mlp.shared_experts.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.6.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.0.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.1.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.2.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.3.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.4.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.5.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.58.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.59.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.6.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.60.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.61.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.62.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.63.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.7.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.router.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.mlp.shared_experts.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.7.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.input_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.0.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.1.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.2.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.3.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.4.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.5.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.58.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.59.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.6.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.60.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.61.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.62.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.63.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.7.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.router.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.mlp.shared_experts.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.post_attention_layernorm.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.8.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.input_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.0.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.1.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.10.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.11.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.12.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.13.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.14.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.15.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.16.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.17.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.18.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.19.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.2.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.20.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.21.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.22.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.23.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.24.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.25.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.26.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.27.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.28.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.29.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.3.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.30.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.31.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.32.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.33.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.34.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.35.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.36.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.37.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.38.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.39.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.4.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.40.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.41.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.42.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.43.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.44.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.45.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.46.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.47.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.48.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.49.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.5.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.50.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.51.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.52.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.53.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.54.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.55.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.56.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.57.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.58.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.59.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.6.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.60.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.61.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.62.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.63.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.7.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.8.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.down_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.down_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.gate_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.gate_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.up_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.experts.experts.9.up_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.router.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.down_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.down_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.gate_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.gate_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.up_proj.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.mlp.shared_experts.up_proj.weight_scale": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.post_attention_layernorm.weight": "model-00003-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.k_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.k_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.o_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.o_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.q_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.q_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.v_proj.weight": "model-00002-of-00006.safetensors",
+    "language_model.model.layers.9.self_attn.v_proj.weight_scale": "model-00002-of-00006.safetensors",
+    "language_model.model.norm.weight": "model-00006-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.layer_norm.bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.layer_norm.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.linear.bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.linear.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.ln_kv.bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.ln_kv.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.multihead_attn.in_proj_bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.multihead_attn.in_proj_weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.multihead_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.multihead_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.cross_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.ffn.linear_in.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.ffn.linear_out.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.ln_ffn.bias": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.ln_ffn.weight": "model-00001-of-00006.safetensors",
+    "multi_modal_projector.query": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.embeddings.patch_embedding.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.embeddings.patch_embedding.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.embeddings.position_embedding.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.0.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.1.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.10.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.11.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.12.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.13.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.14.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.15.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.16.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.17.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.18.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.19.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.2.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.20.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.21.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.22.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.23.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.24.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.25.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.26.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.3.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.4.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.5.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.6.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.7.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.8.self_attn.v_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.layer_norm1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.layer_norm1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.layer_norm2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.layer_norm2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc1.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc1.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc2.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.mlp.fc2.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.k_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.out_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.q_proj.weight": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.bias": "model-00001-of-00006.safetensors",
+    "vision_tower.vision_model.encoder.layers.9.self_attn.v_proj.weight": "model-00001-of-00006.safetensors"
+  }
+}