First model version

config.json

@@ -0,0 +1,23 @@
+{
+  "_name_or_path": "decapoda-research/llama-7b-hf",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "bos_token_id": 0,
+  "eos_token_id": 1,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_sequence_length": 2048,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "pad_token_id": -1,
+  "rms_norm_eps": 1e-06,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.28.0.dev0",
+  "use_cache": true,
+  "vocab_size": 32000
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 0,
+  "transformers_version": "4.28.0.dev0"
+}

pytorch_model-00001-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe0ef23c11b8a309d88bb9e85402a7d60a00dc8f1bb4dee0b3ea6b25e6b8271b
+size 396364479

pytorch_model-00002-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d930e11e9b87132732512f4d24b9f4bfb05b274c3702cb5192292f3ca25d7536
+size 371215393

pytorch_model-00003-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea92b73e1869288754b87803d8e6bf57d1ffd5a49765f53a5be3a3f21d8726be
+size 371215986

pytorch_model-00004-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a762a57267687e1a10fd7af69d36ffbe18110fb9f1933ed2740d1c8a4e58d3b
+size 371215986

pytorch_model-00005-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f348d33d84a918daa275e315738fce9db8ae15c5b2e10ef879e343c0245fe012
+size 371215986

pytorch_model-00006-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:301ae7e2a94b99ac1658d010271d5b0c4d0c1058c588a308f4dfae930ac7b0d1
+size 314575888

pytorch_model-00007-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e51399a151b0870eb1228cdd74205b4480cbb207db3c82027890909a72fa6283
+size 314592882

pytorch_model-00008-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0936d0fc29ff2829fd8c05797221fde0ca2be0b275c1b7439c14e41e1b497a55
+size 314592882

pytorch_model-00009-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a47e74af42361b9ac3cf7bc99c7729945da2d77552834b02ab8738454913df67
+size 371215393

pytorch_model-00010-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b1eec31c32fe0d78eb817817b10195a6c3c76d860077e260e481e4ed77568ed
+size 371215986

pytorch_model-00011-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3aeb135ca0d530a59405af3fa17cacdcb9db22b691ebf344f6942795b7579b2a
+size 371215986

pytorch_model-00012-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:426fdea7900f3594f0ca1c4c43d57b28387e335ef1b0a3ea08351b0d3ac013ea
+size 371215986

pytorch_model-00013-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc747bf5a81a76d74b2bb9793fb19c536409123011070f26aafe139b247e0a5a
+size 314575888

pytorch_model-00014-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0856d9db49d43969aab7681f1f69a9a84f884074b119f5997a77e263f01be7c4
+size 314592882

pytorch_model-00015-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80c8e0bae56d457e993f20911b55e2cd76a1372664dee0121cb93e9c2678726e
+size 314592882

pytorch_model-00016-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07ffaf5be7b3fffa69fd6cf77cafb35faba4e9e94282d57724fc47b0203d4a92
+size 371215393

pytorch_model-00017-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:491f5beb6ef4c6342e648c6efbf85c9f3747cad1eca80e6daeb17c14fb99c9a8
+size 371215986

pytorch_model-00018-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c0e563383a89518495f026d1b272a4068c077ca944e454aa8dd8c385677fe73
+size 371215986

pytorch_model-00019-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:772c6ef37197e622a3a48435f6f062096737e02108723ff124a8b28582cfb61b
+size 371215986

pytorch_model-00020-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5a4b0c77f511a042de17ff1b3d8acdd20de5d25a1ce3e0f75099f0f73f8d61a
+size 314575888

pytorch_model-00021-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddcc405dcf95fd6848b923102698a1524a37dfa64ec862972509ee16e071e2f7
+size 314592882

pytorch_model-00022-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d61580ef6bcc6577a04020869c6a467200d4dd93074f354b098598e97d6617b2
+size 314592882

pytorch_model-00023-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ae55495f61897d0b064493b0282e3e380946917d91fdcdac9d9dc3a8f5a89596
+size 371215393

pytorch_model-00024-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08329a7fd0c10a7052c54367e610bbd06926b41b2cb4a874ffd098da44821f1c
+size 371215986

pytorch_model-00025-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe67453f698af9c39ef71bccae942fd2bec92aebb6129e974129447ff71adfdc
+size 371215986

pytorch_model-00026-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1842139a723a57831c4054a9ec87b9a2d3b5180701fef0ac6ac4f2e9707df53
+size 371215986

pytorch_model-00027-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9cb3bc237248d9e612f77603297d9ef8de8abf1bebd0aaf7d481093917a180b
+size 314575888

pytorch_model-00028-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b9cdc08cb7db2eb2982acee46790529f334c30e750cfc3a3dd575a299f0c7dd
+size 314592882

pytorch_model-00029-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44ef6dd2b61ae21a7958891103c0f65b73347f98627fdaa874183b715e837eaf
+size 314592882

pytorch_model-00030-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24f082831df7294a6e218ef99c95ec5217f97bba4150cd00c6e7f5e0c1620f75
+size 371215393

pytorch_model-00031-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3df293c7311f6b1b7f64a1322204badba8c4be455758ea8a406caa509731266
+size 371215986

pytorch_model-00032-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd64f1768bd68193393e4c6c37c9cbd725044a7e0eb1ac6f3c64f79b7749bd1d
+size 371215986

pytorch_model-00033-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:402f1ef08817f93ab5363b252d1e87be6fba4a352292365786f3c29921427fb2
+size 371215986

pytorch_model-00034-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2798531b22c68475eaef32772e59c5d8d96d7a437aa81d3c632d5947fa31ab0
+size 314575888

pytorch_model-00035-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:faac2917d99c90d9dcbffe27238bb644e84520f61edfdaa0285dc1aa0be6ca52
+size 314592882

pytorch_model-00036-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f5a0e39d0e4651d0022e537116cd6937d270c0e1fb79d9e3fa9bfc41e030e45
+size 314592882

pytorch_model-00037-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45b189388471180c2bc2eb99055a348f82d3bba8d931f58f08c27b74715b2278
+size 371215393

pytorch_model-00038-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6762c675eb30868665ce8ada7a2bbbd117e02f44d47ba3c8dcf91f5623e5ce67
+size 304114977

pytorch_model-00039-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22c5756c5600111814c33f7fcbc9ebbaeadd846ab8ec6f73e6973a1ceb428f31
+size 262144938

pytorch_model.bin.index.json

@@ -0,0 +1,330 @@
+{
+  "metadata": {
+    "total_size": 13476839424
+  },
+  "weight_map": {
+    "base_model.model.lm_head.weight": "pytorch_model-00039-of-00039.bin",
+    "model.embed_tokens.weight": "pytorch_model-00001-of-00039.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00039.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00039.bin",
+    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00039.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00039.bin",
+    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00039.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00039.bin",
+    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00002-of-00039.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00039.bin",
+    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00013-of-00039.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00039.bin",
+    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00014-of-00039.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00039.bin",
+    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00015-of-00039.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00039.bin",
+    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00016-of-00039.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00017-of-00039.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00039.bin",
+    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00018-of-00039.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00039.bin",
+    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00019-of-00039.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00039.bin",
+    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00020-of-00039.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00039.bin",
+    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00021-of-00039.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00039.bin",
+    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00022-of-00039.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00039.bin",
+    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00023-of-00039.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00003-of-00039.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00039.bin",
+    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00024-of-00039.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00039.bin",
+    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00025-of-00039.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00039.bin",
+    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00026-of-00039.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00039.bin",
+    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00027-of-00039.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00039.bin",
+    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00028-of-00039.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00039.bin",
+    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00029-of-00039.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00039.bin",
+    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00030-of-00039.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00031-of-00039.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00039.bin",
+    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00032-of-00039.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00033-of-00039.bin",
+    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00033-of-00039.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00034-of-00039.bin",
+    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00034-of-00039.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00035-of-00039.bin",
+    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00035-of-00039.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00004-of-00039.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00039.bin",
+    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00005-of-00039.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00036-of-00039.bin",
+    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00036-of-00039.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00038-of-00039.bin",
+    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00037-of-00039.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00039.bin",
+    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00006-of-00039.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00039.bin",
+    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00007-of-00039.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00039.bin",
+    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00008-of-00039.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00039.bin",
+    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00009-of-00039.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00010-of-00039.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00039.bin",
+    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00011-of-00039.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00039.bin",
+    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00012-of-00039.bin",
+    "model.norm.weight": "pytorch_model-00038-of-00039.bin"
+  }
+}