Upload BaichuanForCausalLM

config.json

@@ -0,0 +1,28 @@
+{
+  "_from_model_config": true,
+  "_name_or_path": "sharpbai/Baichuan-13B-Base",
+  "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 +1,7 @@
+{
+  "_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-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 sha256:2f1927f5bcc13d978ce01cf29f779dfb728e8c7cf5641d05ca7f54cd60aeca00
+size 655360938

pytorch_model-00003-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cc2940e0b43fc2f126cce50d74a8c9e7a30cc635a88d99db56ad1907c6ca018
+size 630479632

pytorch_model-00004-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cfdbdebf98e88c71c5060e6eb04b3529248175f3ca3df8e75d5cb701ee8cdc9
+size 630479632

pytorch_model-00005-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ae7993ffa2c457c96af2d6333284efca5f85ba9d1f9045f9961724d4089136b
+size 630479632

pytorch_model-00006-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7458a90c0e56bbdd2e30874aba8e5149906c7f2459b95cafde37bd0cc0e813d5
+size 630479632

pytorch_model-00007-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:423d18320d9b79f0f797d364d371c285a87b728b9143979688d4333b2146cb76
+size 630479632

pytorch_model-00008-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:850e4be644f199c1fa11bd42c711502bb25ed9d5cab69df9ca36dd64ae93736d
+size 630479632

pytorch_model-00009-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa7199ba8175d9aed570b58e4acc899c152db235fa503e4ec050557ab4c57da5
+size 630479632

pytorch_model-00010-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfd9c980b38835d785182eb1eb6c4de175470e9c0bec88b6ac29a0b4bf16559c
+size 630479632

pytorch_model-00011-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51c30c2ea5a3bb4d555c28a5a3172a1a128b921c4b03271fcbce18722085693b
+size 630479632

pytorch_model-00012-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c1aed683bc14412d0edae4692f37c71efc5d0e97c5733d46140145600b27c1f
+size 630479632

pytorch_model-00013-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f073df14ce12784eded4791f5a812215a86b9cd6c9986bd713715e954951dc3a
+size 630479632

pytorch_model-00014-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f3244f8ffdb1e2d3089f9a926b5f9857c0f93a2f9652f6d78de620333099fd9
+size 630479632

pytorch_model-00015-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1125bba4f4337db53a1b7ecc7c74118ad885d34b010f37878729e6a670feb51a
+size 630479632

pytorch_model-00016-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:235a476cff156d852253335ead252bdfec2c3bf4a886e087fbe4d07bef84c842
+size 630479632

pytorch_model-00017-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:094806bb593292729d0479a232324a36e5d4042b1b1208cd96d80ec18ffd3996
+size 630479632

pytorch_model-00018-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:849b02e954865f5477822f0c47f6b906bcc8304acc66ce304726af5ac1f20082
+size 630479632

pytorch_model-00019-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14c12f9becd0f82877d034fead1e5b66567a73fbcb455ff6a45d8e346bf6e277
+size 630479632

pytorch_model-00020-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8b5c4ccbb9437d9848a0b92189db10ff12282b945b9472e6de16ab4f5b544a9
+size 630479632

pytorch_model-00021-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3856c2216d435ac5835e916a1608ee3a41a28ead12745afe33f218d25241eaac
+size 630479632

pytorch_model-00022-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8403bc48e4609778d0455e5ac1c1c179691abfd1bbceacc43c05ed9bd7b763a9
+size 630479632

pytorch_model-00023-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85f3e566316848a84a544f393b28e5256c0626253f71a6c8a49eb6bf6099732a
+size 630479632

pytorch_model-00024-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1448515a414b33bdb079903a17a9e53e22647cfb252dd3daa8d32851c6c8b67e
+size 630479632

pytorch_model-00025-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3413796d18a1b828aa6429934185fef3aef4708fa0146b5eeaa4a3351c7140d6
+size 630479632

pytorch_model-00026-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7848c26ec606280136c88c17c1a3962a4f5848a328e5a38f797aa237d159eeb
+size 630479632

pytorch_model-00027-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5908f680a3d707d35665889b5c6cf538807079632faa4a539eb40af8d8bd3d90
+size 630479632

pytorch_model-00028-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3eff6d4d1e3395f973994a3588c4416bd99afca9410d1bfa89400538cf1fc9e5
+size 630479632

pytorch_model-00029-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:081f7190b1ccbf957abe46407caccc0fc2c5c76de69c4eece28dfee7653a3330
+size 630479632

pytorch_model-00030-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b25529c6b62cb3a2c8179be3754f588121966c1578d452672de4a5a8977feec6
+size 630479632

pytorch_model-00031-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1330ca22cc9a41f998462b6f8c69d3cdb58a0da7c1bbc26304fb72470df455b9
+size 630479632

pytorch_model-00032-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ed610253f9ec9dc9a37e420f5f6f4c3aef47342c67138535457b27b35b06996
+size 630479632

pytorch_model-00033-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:978eda6455ecadeca031b779c09d87e94f4abf4db96e1896a50cbb5729b4a76a
+size 630479632

pytorch_model-00034-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed2260ebebfca40635d876efe51932b5b2e189e7b38ca9c033b1727a81845bd4
+size 630479632

pytorch_model-00035-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8606ad2df32a49780151fcb5934099c0faab57bff6a3e8553f57d72af62c3376
+size 630479632

pytorch_model-00036-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2463ec4741036a8c8f3604b762a5bb2adf0797c56cd2da3f06565c0cc6efb64
+size 630479632

pytorch_model-00037-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:740d5a97af71186e6b334bc45b2b7d07768fb2bd0551e5102bdf7deb308ffc88
+size 630479632

pytorch_model-00038-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f58063d6f98f3d9c029affc1f274ecaa1ac735cd7c28de9575d2dd56a829c2a
+size 630479632

pytorch_model-00039-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6775856d6927209ec985f3a08d15cb553c785ae17d039bf4c27b4b4af96a8aa
+size 630479632

pytorch_model-00040-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86f1a7a6b788d340682826650179ca07da77e3ec2de4945d5d8316e4f0643b70
+size 630479632

pytorch_model-00041-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fd51b9ba85a4df16acecedf787d29ba0fa266b93755c75307f26efd78d6ea9a
+size 630479632

pytorch_model-00042-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c18e80d591c3a942feea70bdf819f07732cbbaa316cd843bf3b1e891b758dc3c
+size 630490145

pytorch_model-00043-of-00043.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92a881b8fbf93d093eba37e8a393bc281bd866c4623f0c8966af07ec1e23c28d
+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"
+  }
+}