Initial Version

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "Adun/openthaigpt-1.0.0-beta-7b-ckpt-hf",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 4096,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 0,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.28.0",
+  "use_cache": true,
+  "vocab_size": 56554
+}

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": "4.28.0"
+}

pytorch_model-00001-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01907711dd7ff2d3b41140a8db70515479b64ee6bfff82bade4c876d7f099ee5
+size 1012

pytorch_model-00002-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9d78d02e264931d8bca9572894d2b9b6dcc5cd8bf164c337b67e557477e7d1e
+size 463291781

pytorch_model-00003-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0dbdfaae027c1d4f08d9e399f030abb23ce55e0f44fd80cef6ecdc852a37df0
+size 314576427

pytorch_model-00004-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0df4e96e6bb1a22f42f7870e00ee98f7eb8a2ba6e5bd4e8e9f94b7218562e8b8
+size 314593485

pytorch_model-00005-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6489b058ebba0d2873f4d9ce47214a8e05838ef56d1ee0620dc22dd6745662ea
+size 314593485

pytorch_model-00006-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a44304604b0f6d246584a507c5ebbbc6e39bab215951c0c02ea54e775f586736
+size 371215932

pytorch_model-00007-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff2fb36e497c47307cf536cabcf350287d4350f8eeac7e4eefaa1fc77269b190
+size 371216525

pytorch_model-00008-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e026c9c8a241e3129734a8ba4b0589560192735be3192fd8aa5195055d74c39
+size 371216525

pytorch_model-00009-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f892f7d7fb72b5a0a2a357131b11a87653669ff1b74106ad09f5dc770ea66b5b
+size 371216525

pytorch_model-00010-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:477500cc132d9a891b7b6357c74a73c35bfb417c7922cd8125b9131e57ac35e5
+size 314576427

pytorch_model-00011-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4ad492c07933f6464c0789b504fc0ea3e3c5e7a3bb0f22526c37c1af7bc5573
+size 314593485

pytorch_model-00012-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94ee55a33a9ddf752392321a0bacaffef3dc49a8578aeaded989731303cbd294
+size 314593485

pytorch_model-00013-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf05744ba851a31ae5c03bfdb666618258cd7b34530bc665311abf3394a00cb1
+size 371215932

pytorch_model-00014-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0f133502317d01c63b8349ef6ab24c918de5c52cfcf17317ebdc165869fc519
+size 371216589

pytorch_model-00015-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ef9ae32f8ebab45cf4e64be58919f076788870852048d488d6321a453cc6706
+size 371216589

pytorch_model-00016-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1623526b0eec05ea9ee5dac14e0721e039bfbce2e29101af903cb79d8e1a2202
+size 371216589

pytorch_model-00017-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b83ce6a21a8097432e3768170d8a127bf4f3c3f9dddd9a249ebf55ead6a53ec4
+size 314576427

pytorch_model-00018-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6f10bdbfaa8929f72138f2158a8f33ad525dd3ce2ed21a7093eab7a5ec71b4f
+size 314593485

pytorch_model-00019-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:835f5609b9737fc79590573fd4f77296ebc2b7edd35fca3ec3a9c74bca157530
+size 314593485

pytorch_model-00020-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2465597a28a7d5a09c18164097913a4c58545ed1aae5b357249de33c096a48f3
+size 371215996

pytorch_model-00021-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2cfdc728eaff14a8b76984644ef25e22a2ce5e75c7475a3df9f4d1c4db78d6b9
+size 371216589

pytorch_model-00022-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2d78a59e009deb783adf281c92021aec7a768cfd348df40ae378af3ac03915e
+size 371216589

pytorch_model-00023-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80808a9590b193579764b5be8b6e40a5c2b6797eba5006a4d747208196d310cf
+size 371216589

pytorch_model-00024-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0562a1fecc2fef650b4909437faace2bba5deec87082238302196869a4449283
+size 314576427

pytorch_model-00025-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6ce2df0109a1f9e167faadb459e6b03c836f995d3faebf3437374188cd53352
+size 314593485

pytorch_model-00026-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82d11a20439f8dd6314a85004888b7c5426853f7dbea45e61cc71a61dd9009f3
+size 314593485

pytorch_model-00027-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:735aa87b41e30a4903e7c549a6678afe5f31c8aaf9aea4a7971c6da637b33cbd
+size 371215996

pytorch_model-00028-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4f668ad29904213af17858c6ddf0e2a8dbca005aa0877defe9cce83ad7a9b02
+size 371216589

pytorch_model-00029-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5667c12eef2defaba8397fbe8b4207f20a85edc1897cce43e48858d81e3e4fed
+size 371216589

pytorch_model-00030-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfde8a7b72178a9ad0b221e949765bbb0e9513211b90d6a185a74f02e0d0a78d
+size 371216589

pytorch_model-00031-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:baa882b6fcade851b7915d3142f1acef74ec66a9f3ee107adf7932512ac514b4
+size 314576427

pytorch_model-00032-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4471d7be8f788b4278f8a17685fa4ab10ef29ea1af796c8d5af90fbbe79e555
+size 314593485

pytorch_model-00033-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c649b3133f508a25e18ebe98f0185964cfe9895b98a3bdee2a5989a398d84f77
+size 314593485

pytorch_model-00034-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f86df1a74084b054ba047955aecd44c356cedbd183a343be4a7e30719ec617ff
+size 371215996

pytorch_model-00035-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:805a3f56ab8f6984dd603409b8dda657933b55cba90d2996842c03e1e5e900d0
+size 371216589

pytorch_model-00036-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a70688e54a8071adca948232bdc29db90053604025a38c943ed28439fb07d448
+size 371216589

pytorch_model-00037-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5dfcd0c639f672c2f58263e23bce344c0339976b4dfeee0b1b14848aecd804b
+size 371216589

pytorch_model-00038-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b94067b2e33633f88ce8c5b0c2d56578329b328aeb9191f07fc8efb7f2a2721b
+size 314576427

pytorch_model-00039-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:292bdb3f449c899e941fa0d3b6926e4b21b51b84f437e2a0723e62c4cd835089
+size 314593485

pytorch_model-00040-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8d88dbd2118e5b909baf85428e8a758add045cbbff3dc528eeffeed5cd8d5ff
+size 180382345

pytorch_model-00041-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1663a024637ae7aba6a0bf077b714f13f0c852dc8c7e31cb7385b5a49310f84
+size 463291781

pytorch_model.bin.index.json

@@ -0,0 +1,330 @@
+{
+  "metadata": {
+    "total_size": 13879132160
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00041-of-00041.bin",
+    "model.embed_tokens.weight": "pytorch_model-00002-of-00041.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00041.bin",
+    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00003-of-00041.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00041.bin",
+    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00004-of-00041.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00041.bin",
+    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00015-of-00041.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00041.bin",
+    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00016-of-00041.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00041.bin",
+    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00017-of-00041.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00041.bin",
+    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00018-of-00041.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00041.bin",
+    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00019-of-00041.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00041.bin",
+    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00020-of-00041.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00021-of-00041.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00041.bin",
+    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00022-of-00041.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00041.bin",
+    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00023-of-00041.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00041.bin",
+    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00024-of-00041.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00041.bin",
+    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00025-of-00041.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00041.bin",
+    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00005-of-00041.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00041.bin",
+    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00026-of-00041.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00041.bin",
+    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00027-of-00041.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00028-of-00041.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00041.bin",
+    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00029-of-00041.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00041.bin",
+    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00030-of-00041.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00041.bin",
+    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00031-of-00041.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00041.bin",
+    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00032-of-00041.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00033-of-00041.bin",
+    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00033-of-00041.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00035-of-00041.bin",
+    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00034-of-00041.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00035-of-00041.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00036-of-00041.bin",
+    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00036-of-00041.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00037-of-00041.bin",
+    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00037-of-00041.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00041.bin",
+    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00006-of-00041.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00038-of-00041.bin",
+    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00038-of-00041.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00040-of-00041.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00040-of-00041.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00040-of-00041.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00040-of-00041.bin",
+    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00039-of-00041.bin",
+    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00039-of-00041.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00007-of-00041.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00041.bin",
+    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00008-of-00041.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00041.bin",
+    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00009-of-00041.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00041.bin",
+    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00010-of-00041.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00041.bin",
+    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00011-of-00041.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00041.bin",
+    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00012-of-00041.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00041.bin",
+    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00013-of-00041.bin",
+    "model.norm.weight": "pytorch_model-00040-of-00041.bin"
+  }
+}