Upload BaiChuanForCausalLM

config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "sharpbai/Baichuan-7B",
+  "architectures": [
+    "BaiChuanForCausalLM"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_baichuan.BaiChuanConfig",
+    "AutoModelForCausalLM": "modeling_baichuan.BaiChuanForCausalLM"
+  },
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "baichuan",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.30.2",
+  "use_cache": true,
+  "vocab_size": 64000
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.30.2"
+}

pytorch_model-00001-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf02c4d7b83fa495ef73f290168a195ea899d0744f0fe32194046846bd87efe0
+size 537

pytorch_model-00002-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3aff8e9aa72626b49788fd10b0190a35779e5443c7571355b81f937320c06b10
+size 524288938

pytorch_model-00003-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:165a1b52f07880508d750d2ea689a848fbdfe6823d959fa9ab3f8509c15d2c88
+size 404770081

pytorch_model-00004-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:257a98b8e1a90d7cbefe525f2d4a7f44206f529e885efaac425eba25af995318
+size 404770081

pytorch_model-00005-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63cec13ae8b6c09386f622ef8b99d6309652bfa57c2b4baa769d8a1a664e07f2
+size 404770081

pytorch_model-00006-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75904899e0507e7ba4717f59bb81dc327f335b0abff5f1579d9f754dcbbb518c
+size 404770081

pytorch_model-00007-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03e9e6fceed54c314c311982cc12f4b6b0fa14f108dc6d3575bd10afc7f6243f
+size 404770081

pytorch_model-00008-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28a8f678eb6089a63559eb758ba67dd2760f9206546141eb6f77c39a37a77d4d
+size 404770081

pytorch_model-00009-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8eae231300f967af5d710f75f779792f1f65c24d527ad5b8ee76111ca40a177
+size 404770081

pytorch_model-00010-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c54bdee8c8481c622cee6126f13b80d8c1d8a4b8166a2c973d9f1cde9f3e69c4
+size 404770081

pytorch_model-00011-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69c7a5d31b6609971c6fbc956ea81158c2fd49b93f4b8e8856bcba29fc949409
+size 404770081

pytorch_model-00012-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b63565001d8d760e011240349d5a10057c96543bdc54c09459b6e12485faaf91
+size 404770081

pytorch_model-00013-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11aa46a9512bc6a375ff7c9af5482e2390b8d2383dbdc3a974b87f8f2c209462
+size 404770145

pytorch_model-00014-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bfd4b1e3545bff3527b47924e8b89befe73434c62cff503032f81c13a01b5cbe
+size 404770145

pytorch_model-00015-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c618f43e9e4eb03231cb9d49a0b8bd083f905e3bb2cbf4134c5f4b69367b40ad
+size 404770145

pytorch_model-00016-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fc8fb7508d8471a6221d9d40014725f3eea3007bb5ec25de48f0af3126108f7
+size 404770145

pytorch_model-00017-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:240e7eeeb6fcdd5f7abfd019216eea0e0273fff7f9182f86b3bb842a4bc088f8
+size 404770145

pytorch_model-00018-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e2461005e9692766cba28e9763e6a62594c9fa0f5cfc6e28f60f0aa1677f90d
+size 404770145

pytorch_model-00019-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4383b171cc84b015d54450cdee3a7c862ebb701a08c86fa1d824ce98f5db3e9
+size 404770145

pytorch_model-00020-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc111534af81249f501236ab2879f195f2c06007a6c8dca9333abf3c0d3ed476
+size 404770145

pytorch_model-00021-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25f027f668eb0e0520e434a9bf22f0b947384a44b842fa32a63feeb50a915687
+size 404770145

pytorch_model-00022-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c3c5571d868e636e02e293bbde43da179ea8aecab2deec4862b473bc619195e
+size 404770145

pytorch_model-00023-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93c520e0cc022703c2d2df2b3e3cb550e9d43956a3e40359c0cf176c00e8f1fa
+size 404770145

pytorch_model-00024-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbd0502f9538f4b8362dbce4dce743bf3ff536ee292c2185f39c2962deefe21c
+size 404770145

pytorch_model-00025-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c5c94073da5639f4b7033188a4861196268fbd0c5d12f6639c63f1d737cd5ea
+size 404770145

pytorch_model-00026-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41fec1041e1fd03ac77c318148845c665ab81b1a437c193686ce1d52e2baf922
+size 404770145

pytorch_model-00027-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fc3039c09d4cb06e7ecf60011928cccbce65fab5fbb612e3a692b48f7b0e0a4
+size 404770145

pytorch_model-00028-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3de19a551675ec2bb125c0cd006345b17bf82ca25c89170d1e9a15db98b045d
+size 404770145

pytorch_model-00029-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0dc7dd8842cd59768585ff29d66be2c8e1eecfa1fdc2f5e5aacf3fff83cd868
+size 404770145

pytorch_model-00030-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d1be928100f85b7538f7bba9706e1b9eabd1de9de2253ee511133b0434453b3
+size 404770145

pytorch_model-00031-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67abf02ec2bca60b6e18351627368f312294c18bd9f34be33ce922d8bc706c97
+size 404770145

pytorch_model-00032-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cc51fc0d9a5c195b52cd83a8ae00029dfb4a8b6e5724cf191cdbd5f76e4cec7
+size 404770145

pytorch_model-00033-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd4b51f1158d24f8ec33413e3a01bb415db923ee4a9fcac42a7ee07428199099
+size 404770145

pytorch_model-00034-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c1eab4cdbb0b00cab806b78753c35c1f56a7cb6f64d17a196fb61fe005df384
+size 404778610

pytorch_model-00035-of-00035.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:587eaed49ee9e8ca1e5dac03ee8621ddfc3d3a7d1eae18e895d5134fedb079ca
+size 524288938

pytorch_model.bin.index.json

@@ -0,0 +1,266 @@
+{
+  "metadata": {
+    "total_size": 14001127424
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00035-of-00035.bin",
+    "model.embed_tokens.weight": "pytorch_model-00002-of-00035.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.self_attn.W_pack.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00003-of-00035.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00035.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.self_attn.W_pack.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00004-of-00035.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00035.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.self_attn.W_pack.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00013-of-00035.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00035.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.self_attn.W_pack.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00014-of-00035.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00035.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.self_attn.W_pack.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00015-of-00035.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00035.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.self_attn.W_pack.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00016-of-00035.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00035.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.self_attn.W_pack.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00017-of-00035.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00035.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.self_attn.W_pack.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00018-of-00035.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00035.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.self_attn.W_pack.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00019-of-00035.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00035.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.self_attn.W_pack.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00020-of-00035.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00035.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.self_attn.W_pack.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00021-of-00035.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00035.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.self_attn.W_pack.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00022-of-00035.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00035.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.self_attn.W_pack.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00005-of-00035.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00035.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.self_attn.W_pack.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00023-of-00035.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00035.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.self_attn.W_pack.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00024-of-00035.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00035.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.self_attn.W_pack.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00025-of-00035.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00035.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.self_attn.W_pack.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00026-of-00035.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00035.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.self_attn.W_pack.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00027-of-00035.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00035.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.self_attn.W_pack.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00028-of-00035.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00035.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.self_attn.W_pack.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00029-of-00035.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00035.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.self_attn.W_pack.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00030-of-00035.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00035.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.self_attn.W_pack.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00031-of-00035.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00035.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.self_attn.W_pack.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00032-of-00035.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00035.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.self_attn.W_pack.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00006-of-00035.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00035.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.self_attn.W_pack.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00033-of-00035.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00033-of-00035.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.self_attn.W_pack.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00034-of-00035.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00034-of-00035.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.self_attn.W_pack.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00007-of-00035.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00035.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.self_attn.W_pack.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00008-of-00035.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00035.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.self_attn.W_pack.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00009-of-00035.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00035.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.self_attn.W_pack.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00010-of-00035.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00035.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.self_attn.W_pack.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00011-of-00035.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00035.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.self_attn.W_pack.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00012-of-00035.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00035.bin",
+    "model.norm.weight": "pytorch_model-00034-of-00035.bin"
+  }
+}