Upload MPTForCausalLM

config.json

@@ -0,0 +1,53 @@
+{
+  "_name_or_path": "./mpt-30b-orca-hf/",
+  "architectures": [
+    "MPTForCausalLM"
+  ],
+  "attn_config": {
+    "alibi": true,
+    "alibi_bias_max": 8,
+    "attn_impl": "torch",
+    "attn_pdrop": 0,
+    "attn_type": "multihead_attention",
+    "attn_uses_sequence_id": false,
+    "clip_qkv": null,
+    "prefix_lm": false,
+    "qk_ln": false,
+    "softmax_scale": null
+  },
+  "auto_map": {
+    "AutoConfig": "configuration_mpt.MPTConfig",
+    "AutoModelForCausalLM": "modeling_mpt.MPTForCausalLM"
+  },
+  "d_model": 7168,
+  "emb_pdrop": 0,
+  "embedding_fraction": 1.0,
+  "expansion_ratio": 4,
+  "init_config": {
+    "emb_init_std": null,
+    "emb_init_uniform_lim": null,
+    "fan_mode": "fan_in",
+    "init_div_is_residual": true,
+    "init_gain": 0.0,
+    "init_nonlinearity": "relu",
+    "init_std": null,
+    "name": "kaiming_normal_",
+    "verbose": 0
+  },
+  "init_device": "cpu",
+  "learned_pos_emb": true,
+  "logit_scale": null,
+  "max_seq_len": 8192,
+  "model_type": "mpt",
+  "n_heads": 64,
+  "n_layers": 48,
+  "no_bias": true,
+  "norm_type": "low_precision_layernorm",
+  "resid_pdrop": 0,
+  "tokenizer_name": "sam-mosaic/gpt-neox-20b-chatml",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.30.2",
+  "use_cache": false,
+  "verbose": 0,
+  "vocab_size": 50432
+}

generation_config.json

@@ -0,0 +1,5 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.30.2",
+  "use_cache": false
+}

pytorch_model-00001-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c588c385150af7eb487bbbcb0790b95ab94aae85c6205176c541ec4955268bb1
+size 9766157965

pytorch_model-00002-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87437a65e59d6de4600c95cb332be8b822589ebd4e92e26cc08a6488134090f7
+size 9865248775

pytorch_model-00003-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dadf2be1fe567191ab04f02afe81aa094c2a0f6369e66c72886cfe318d2106d
+size 9865248775

pytorch_model-00004-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:196d4a7980737577ee9be3a9eaac6edebdab3b217514d8ce97d3eb6a44cc3101
+size 9865248775

pytorch_model-00005-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeab2a63765c44c9bbae5725cfd32e7e34786d157e3c7c520625b19c9d20aaec
+size 9865248775

pytorch_model-00006-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd2c8d3f360b6fd0d140a93b93d4ef5065d0e877c272fcbad14c7f1dc77556dc
+size 9865248775

pytorch_model-00007-of-00007.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b93e2801018848347cfac842930ca62fd9c09e0968acd514526b314b8f9965b
+size 822099468

pytorch_model.bin.index.json

@@ -0,0 +1,297 @@
+{
+  "metadata": {
+    "total_size": 59914401792
+  },
+  "weight_map": {
+    "transformer.blocks.0.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.0.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.0.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.0.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.0.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.0.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.1.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.10.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.10.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.10.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.10.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.10.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.10.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.11.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.12.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.13.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.14.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.15.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.15.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.15.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.15.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.15.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.15.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.16.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.16.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.16.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.16.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.16.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.16.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.17.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.18.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.19.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.2.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.2.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.2.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.2.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.2.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.2.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.20.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.20.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.20.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.20.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.20.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.20.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.21.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.ffn.down_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.ffn.up_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.22.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.23.attn.Wqkv.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.23.attn.out_proj.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.23.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.23.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.23.norm_1.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.23.norm_2.weight": "pytorch_model-00003-of-00007.bin",
+    "transformer.blocks.24.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.24.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.24.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.24.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.24.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.24.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.25.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.26.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.27.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.28.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.29.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.3.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.3.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.3.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.3.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.3.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.3.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.30.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.30.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.30.ffn.down_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.30.ffn.up_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.30.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.30.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.31.attn.Wqkv.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.31.attn.out_proj.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.31.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.31.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.31.norm_1.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.31.norm_2.weight": "pytorch_model-00004-of-00007.bin",
+    "transformer.blocks.32.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.32.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.32.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.32.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.32.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.32.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.33.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.34.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.35.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.36.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.37.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.ffn.down_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.ffn.up_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.38.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.39.attn.Wqkv.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.39.attn.out_proj.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.39.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.39.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.39.norm_1.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.39.norm_2.weight": "pytorch_model-00005-of-00007.bin",
+    "transformer.blocks.4.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.4.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.4.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.4.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.4.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.4.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.40.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.40.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.40.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.40.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.40.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.40.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.41.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.42.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.43.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.44.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.45.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.ffn.down_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.ffn.up_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.46.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.47.attn.Wqkv.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.47.attn.out_proj.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.47.ffn.down_proj.weight": "pytorch_model-00007-of-00007.bin",
+    "transformer.blocks.47.ffn.up_proj.weight": "pytorch_model-00007-of-00007.bin",
+    "transformer.blocks.47.norm_1.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.47.norm_2.weight": "pytorch_model-00006-of-00007.bin",
+    "transformer.blocks.5.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.5.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.5.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.5.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.5.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.5.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.ffn.down_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.ffn.up_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.6.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.7.attn.Wqkv.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.7.attn.out_proj.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.7.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.7.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.7.norm_1.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.7.norm_2.weight": "pytorch_model-00001-of-00007.bin",
+    "transformer.blocks.8.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.8.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.8.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.8.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.8.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.8.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.attn.Wqkv.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.attn.out_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.ffn.down_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.ffn.up_proj.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.norm_1.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.blocks.9.norm_2.weight": "pytorch_model-00002-of-00007.bin",
+    "transformer.norm_f.weight": "pytorch_model-00007-of-00007.bin",
+    "transformer.wte.weight": "pytorch_model-00001-of-00007.bin"
+  }
+}