Upload BaichuanForCausalLM

config.json

@@ -0,0 +1,28 @@
+{
+  "_from_model_config": true,
+  "_name_or_path": "sharpbai/Baichuan-13B-Chat",
+  "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": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 13696,
+  "model_max_length": 4096,
+  "model_type": "baichuan",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 40,
+  "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 +1,14 @@
+{
+  "assistant_token_id": 196,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "max_new_tokens": 2048,
+  "pad_token_id": 0,
+  "repetition_penalty": 1.1,
+  "temperature": 0.3,
+  "top_k": 5,
+  "top_p": 0.85,
+  "transformers_version": "4.30.2",
+  "user_token_id": 195
+}

pytorch_model-00001-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10f795bc1b0392a20c7548df7092bd5897e8f6fd51bfb013a471fdf018374082
+size 537

pytorch_model-00002-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7bfc6b9aee818f931017a0b6c9b9e66f0a44af6db477a1fc8dc69e708d17c756
+size 655360938

pytorch_model-00003-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e9531931246eecee4ecca7ed66cb5743dc3992ba6c61f072d4a77489f9d49bd
+size 630479632

pytorch_model-00004-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b51d79b625f41cda3b9754272bd4addeca8df204f513a4d06237c5eb1c1b7a3
+size 630479632

pytorch_model-00005-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:950e7cf423e651dd60ce005d0510d9d4eb185a02472e869e76d12d6975f0325f
+size 630479632

pytorch_model-00006-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9e825e43a799baae5c57f49daaafa15730153a470aaf22b5c4e02050bf7c613
+size 630479632

pytorch_model-00007-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8af5281f5ec95174974badc75cfb4c5f047239021e34a676a2693cac59462e8
+size 630479632

pytorch_model-00008-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45af417eae5a6370535eba25fa35369d7b783b085ed923e253fce426c4a2dc24
+size 630479632

pytorch_model-00009-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9597b16fc0b48a4c2d432cbdfc5dcddb61e1ca02b74f81134fb0d8c9ebedbb10
+size 630479632

pytorch_model-00010-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3070df4350aebc3e482eb30381427c28d8c00be0f52f165bafea03019e07eadf
+size 630479632

pytorch_model-00011-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c3e1a94626904ad22d5708d5d4e0f2306d3b7773125bd7dca49eaa457363c52
+size 630479632

pytorch_model-00012-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c6c5bf0c514e97b7afd57f2240081e310c7909e9231fa6e6b5f7eb433b66154
+size 630479632

pytorch_model-00013-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2030ed0f36c8facfe2c6077b3b861acde85e8b62ace46deaf32482696678e857
+size 630479632

pytorch_model-00014-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abd103d1edab58d53b13b43cd1aba69cf373f2be641aa5477e19d424bbe62ca5
+size 630479632

pytorch_model-00015-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:161fad84e95b1ad2c137b11687137a97646c7ac56b730e93485a41fd080a41c6
+size 630479632

pytorch_model-00016-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a213b6238f42b3e460a160c1bd0ce62345b348b6d40a1949cb710237a72fdcd
+size 630479632

pytorch_model-00017-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80f2597098ea4193e7743c020a816aba5c2eed8addcc8b34c0cfda4cc1ba324d
+size 630479632

pytorch_model-00018-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8660c173992c2ec75709f3210679e116f2473d58cbe3a7204095a718d1e4565
+size 630479632

pytorch_model-00019-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9217a1ac05d6a27b683cb3744e6729533539eddb4229dad4a86a10ed056570cb
+size 630479632

pytorch_model-00020-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a5a3797dd859d88517547d418e3075ce51184e3e53ccf8df063681a947d5b86
+size 630479632

pytorch_model-00021-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:49b0ec4f09aeaf19e64ea594d1018a0130ca7fac79c8bb76931a8ee570d9dbd8
+size 630479632

pytorch_model-00022-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3126a1ce34bfa4fbd14b035b29beb0c34baa1345e54bc6ed23cdbc0c3c8d204b
+size 630479632

pytorch_model-00023-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4155f26f39adee65f603e7efb075bfabeb5323bbbe291cda5fa28952203772ff
+size 630479632

pytorch_model-00024-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a83e8819fbd37083b29b48fff508566d5b11099cd7d859667e46c58bbe6a633
+size 630479632

pytorch_model-00025-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a40e93e2b23ec20149c68ee86facf8395ae76664af69015d996c667f17ccee8f
+size 630479632

pytorch_model-00026-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b44145c502b5a72320fe80f50064ac62bf8d66b28937ff78183717c644d9af16
+size 630479632

pytorch_model-00027-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ef8a496bb865782a4f25ce66e142028786a10c4f674988cdc98359160b1dd2c
+size 630479632

pytorch_model-00028-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea246f78707c0f711fea931a4fb3c5039c68cd4d3499b2e1ef4c187d3c111f90
+size 630479632

pytorch_model-00029-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:503ce93c57fb521e95396a94f8bb071f75d315f85767b0980fb3b49493391505
+size 630479632

pytorch_model-00030-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eba0f5d6bcc5807a59837add3786cd297edc218cf732e024c47278d1aefcbced
+size 630479632

pytorch_model-00031-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba2d9fca2ede1ad9c2528b6e80f71dab00910b992abb583761014c3ebe3cd128
+size 630479632

pytorch_model-00032-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf00bdb67f771a2ed770cdc2d69a8cecfe964d7dd8ee48d07ff6c89d6c184531
+size 630479632

pytorch_model-00033-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48f4acd21b6923fbee8e67671adfe7447a9563cf7932fc95e8ce74e1668f859d
+size 630479632

pytorch_model-00034-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70e30826d04c9790ea28c3ffdb3581d73d49f74565220f73cb79bd922fb2e664
+size 630479632

pytorch_model-00035-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c4b056a0efa0440151577bac566bcb73976df2756127ea76dbe636b050c9b6
+size 630479632

pytorch_model-00036-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a889b74a6bc979e112c5c2458afd48e9d1687632403e646b26fde2fa7f6b16c
+size 630479632

pytorch_model-00037-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:727b23dd5b75c29e8a1edb426a9a77546bf096fbef38ac67e1a97f91cb2c7ed5
+size 630479632

pytorch_model-00038-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1bdc612e5a77c6872e79213955176db04dc7bade73db9d49c13ffeb98c10522
+size 630479632

pytorch_model-00039-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08ebe1d7ee8e27d2327240e90c7a141ae05c655da11fd5f5d1c560a89438753a
+size 630479632

pytorch_model-00040-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:882cec6c572ed3628d3842b5b78b1a9cf7614e6cbb0db4e12e8974606de52455
+size 630479632

pytorch_model-00041-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc89a954b3183e9a82bddaa96d2be5d16bb70dd1ac8cb0f183b6d59949057a7f
+size 630479632

pytorch_model-00042-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a42e6810d7394bbd25b82b187d055f0eebca0020c950ca33c65a04ae9deb5cee
+size 630490145

pytorch_model-00043-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bf38e92315430a5c9e00aeb190a1f9177b1dd088ae70003bd2330a862f57d5f
+size 655360938

pytorch_model.bin.index.json

@@ -0,0 +1,290 @@
+{
+  "metadata": {
+    "total_size": 26529802240
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00043-of-00043.bin",
+    "model.embed_tokens.weight": "pytorch_model-00002-of-00043.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.self_attn.W_pack.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00003-of-00043.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.self_attn.W_pack.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00004-of-00043.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.self_attn.W_pack.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00013-of-00043.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.self_attn.W_pack.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00014-of-00043.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.self_attn.W_pack.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00015-of-00043.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.self_attn.W_pack.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00016-of-00043.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.self_attn.W_pack.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00017-of-00043.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.self_attn.W_pack.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00018-of-00043.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.self_attn.W_pack.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00019-of-00043.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.self_attn.W_pack.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00020-of-00043.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.self_attn.W_pack.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00021-of-00043.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.self_attn.W_pack.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00022-of-00043.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.self_attn.W_pack.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00005-of-00043.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.self_attn.W_pack.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00023-of-00043.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.self_attn.W_pack.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00024-of-00043.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.self_attn.W_pack.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00025-of-00043.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.self_attn.W_pack.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00026-of-00043.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.self_attn.W_pack.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00027-of-00043.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.self_attn.W_pack.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00028-of-00043.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.self_attn.W_pack.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00029-of-00043.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.self_attn.W_pack.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00030-of-00043.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.self_attn.W_pack.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00031-of-00043.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.self_attn.W_pack.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00032-of-00043.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.self_attn.W_pack.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00006-of-00043.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.self_attn.W_pack.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00033-of-00043.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.self_attn.W_pack.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00034-of-00043.bin",
+    "model.layers.32.input_layernorm.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.mlp.down_proj.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.mlp.up_proj.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.self_attn.W_pack.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00035-of-00043.bin",
+    "model.layers.33.input_layernorm.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.mlp.down_proj.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.mlp.up_proj.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.self_attn.W_pack.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00036-of-00043.bin",
+    "model.layers.34.input_layernorm.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.mlp.down_proj.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.mlp.up_proj.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.self_attn.W_pack.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00037-of-00043.bin",
+    "model.layers.35.input_layernorm.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.mlp.down_proj.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.mlp.up_proj.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.self_attn.W_pack.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00038-of-00043.bin",
+    "model.layers.36.input_layernorm.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.mlp.down_proj.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.mlp.up_proj.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.self_attn.W_pack.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00039-of-00043.bin",
+    "model.layers.37.input_layernorm.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.mlp.down_proj.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.mlp.up_proj.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.self_attn.W_pack.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00040-of-00043.bin",
+    "model.layers.38.input_layernorm.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.mlp.down_proj.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.mlp.up_proj.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.self_attn.W_pack.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00041-of-00043.bin",
+    "model.layers.39.input_layernorm.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.mlp.down_proj.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.mlp.up_proj.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.self_attn.W_pack.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00042-of-00043.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.self_attn.W_pack.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00007-of-00043.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.self_attn.W_pack.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00008-of-00043.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.self_attn.W_pack.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00009-of-00043.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.self_attn.W_pack.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00010-of-00043.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.self_attn.W_pack.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00011-of-00043.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.self_attn.W_pack.weight": "pytorch_model-00012-of-00043.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00012-of-00043.bin",
+    "model.norm.weight": "pytorch_model-00042-of-00043.bin"
+  }
+}