Upload MixFormerSequentialForCausalLM

config.json

@@ -0,0 +1,30 @@
+{
+  "_name_or_path": "microsoft/phi-1_5",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "MixFormerSequentialForCausalLM"
+  ],
+  "attn_pdrop": 0.0,
+  "auto_map": {
+    "AutoConfig": "microsoft/phi-1_5--configuration_mixformer_sequential.MixFormerSequentialConfig",
+    "AutoModelForCausalLM": "microsoft/phi-1_5--modeling_mixformer_sequential.MixFormerSequentialForCausalLM"
+  },
+  "embd_pdrop": 0.0,
+  "flash_rotary": false,
+  "fused_dense": false,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "mixformer-sequential",
+  "n_embd": 2048,
+  "n_head": 32,
+  "n_head_kv": null,
+  "n_inner": null,
+  "n_layer": 24,
+  "n_positions": 2048,
+  "resid_pdrop": 0.0,
+  "rotary_dim": 32,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.35.0",
+  "vocab_size": 51200
+}

generation_config.json

@@ -0,0 +1,4 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.35.0"
+}

model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13e7443c7d257104101a1bd094efa533ee8e07fb9842a8adfacb80161172f3af
+size 4984920088

model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50b2710418a29044d4c388e36686329146f7f88b86ecb2bec294608dad24a3de
+size 688186896

model.safetensors.index.json

@@ -0,0 +1,252 @@
+{
+  "metadata": {
+    "total_size": 5673082880
+  },
+  "weight_map": {
+    "layers.0.wte.weight": "model-00001-of-00002.safetensors",
+    "layers.1.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.1.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.1.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.1.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.1.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.1.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.1.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.1.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.1.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.1.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.10.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.10.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.10.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.10.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.10.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.10.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.10.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.10.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.10.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.10.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.11.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.11.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.11.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.11.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.11.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.11.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.11.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.11.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.11.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.11.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.12.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.12.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.12.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.12.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.12.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.12.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.12.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.12.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.12.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.12.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.13.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.13.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.13.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.13.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.13.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.13.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.13.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.13.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.13.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.13.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.14.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.14.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.14.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.14.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.14.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.14.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.14.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.14.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.14.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.14.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.15.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.15.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.15.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.15.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.15.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.15.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.15.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.15.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.15.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.15.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.16.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.16.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.16.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.16.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.16.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.16.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.16.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.16.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.16.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.16.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.17.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.17.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.17.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.17.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.17.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.17.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.17.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.17.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.17.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.17.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.18.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.18.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.18.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.18.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.18.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.18.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.18.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.18.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.18.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.18.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.19.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.19.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.19.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.19.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.19.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.19.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.19.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.19.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.19.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.19.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.2.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.2.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.2.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.2.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.2.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.2.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.2.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.2.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.2.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.2.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.20.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.20.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.20.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.20.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.20.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.20.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.20.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.20.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.20.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.20.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.21.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.21.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.21.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.21.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.21.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.21.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.21.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.21.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.21.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.21.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.22.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.22.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.22.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.22.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.22.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.22.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.22.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.22.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.22.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.22.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.23.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.23.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.23.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.23.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.23.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.23.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.23.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.23.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.23.mlp.fc2.bias": "model-00002-of-00002.safetensors",
+    "layers.23.mlp.fc2.weight": "model-00002-of-00002.safetensors",
+    "layers.24.ln.bias": "model-00002-of-00002.safetensors",
+    "layers.24.ln.weight": "model-00002-of-00002.safetensors",
+    "layers.24.mixer.Wqkv.bias": "model-00002-of-00002.safetensors",
+    "layers.24.mixer.Wqkv.weight": "model-00002-of-00002.safetensors",
+    "layers.24.mixer.out_proj.bias": "model-00002-of-00002.safetensors",
+    "layers.24.mixer.out_proj.weight": "model-00002-of-00002.safetensors",
+    "layers.24.mlp.fc1.bias": "model-00002-of-00002.safetensors",
+    "layers.24.mlp.fc1.weight": "model-00002-of-00002.safetensors",
+    "layers.24.mlp.fc2.bias": "model-00002-of-00002.safetensors",
+    "layers.24.mlp.fc2.weight": "model-00002-of-00002.safetensors",
+    "layers.25.linear.bias": "model-00002-of-00002.safetensors",
+    "layers.25.linear.weight": "model-00002-of-00002.safetensors",
+    "layers.25.ln.bias": "model-00002-of-00002.safetensors",
+    "layers.25.ln.weight": "model-00002-of-00002.safetensors",
+    "layers.3.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.3.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.3.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.3.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.3.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.3.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.3.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.3.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.3.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.3.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.4.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.4.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.4.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.4.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.4.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.4.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.4.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.4.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.4.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.4.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.5.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.5.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.5.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.5.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.5.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.5.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.5.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.5.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.5.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.5.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.6.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.6.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.6.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.6.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.6.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.6.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.6.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.6.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.6.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.6.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.7.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.7.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.7.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.7.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.7.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.7.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.7.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.7.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.7.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.7.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.8.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.8.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.8.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.8.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.8.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.8.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.8.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.8.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.8.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.8.mlp.fc2.weight": "model-00001-of-00002.safetensors",
+    "layers.9.ln.bias": "model-00001-of-00002.safetensors",
+    "layers.9.ln.weight": "model-00001-of-00002.safetensors",
+    "layers.9.mixer.Wqkv.bias": "model-00001-of-00002.safetensors",
+    "layers.9.mixer.Wqkv.weight": "model-00001-of-00002.safetensors",
+    "layers.9.mixer.out_proj.bias": "model-00001-of-00002.safetensors",
+    "layers.9.mixer.out_proj.weight": "model-00001-of-00002.safetensors",
+    "layers.9.mlp.fc1.bias": "model-00001-of-00002.safetensors",
+    "layers.9.mlp.fc1.weight": "model-00001-of-00002.safetensors",
+    "layers.9.mlp.fc2.bias": "model-00001-of-00002.safetensors",
+    "layers.9.mlp.fc2.weight": "model-00001-of-00002.safetensors"
+  }
+}