{ "metadata": { "total_size": 10402539520 }, "weight_map": { "transformer.blocks.0.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.0.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.0.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.0.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.0.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.0.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.1.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.10.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.11.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.12.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.13.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.14.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.15.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.16.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.17.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.18.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.19.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.2.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.20.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.21.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.22.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.23.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.24.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.25.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.26.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.27.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.28.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.29.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.3.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.30.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.30.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.30.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.30.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.30.mlp.mlp_down.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.30.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.31.attn.Wqkv.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.31.attn.out_proj.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.31.ln_1.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.31.ln_2.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.31.mlp.mlp_down.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.31.mlp.mlp_up.weight": "pytorch_model-00002-of-00002.bin", "transformer.blocks.4.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.4.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.4.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.4.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.4.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.4.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.5.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.6.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.7.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.8.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.attn.Wqkv.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.attn.out_proj.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.ln_1.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.ln_2.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.mlp.mlp_down.weight": "pytorch_model-00001-of-00002.bin", "transformer.blocks.9.mlp.mlp_up.weight": "pytorch_model-00001-of-00002.bin", "transformer.ln_f.weight": "pytorch_model-00002-of-00002.bin", "transformer.wte.weight": "pytorch_model-00001-of-00002.bin" } }