Upload LlamaForCausalLM

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "sharpbai/Llama-2-7b-chat",
+  "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": 2048,
+  "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": "4.30.2",
+  "use_cache": true,
+  "vocab_size": 32000
+}

generation_config.json

@@ -0,0 +1,9 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 32000,
+  "temperature": 0.9,
+  "top_p": 0.6,
+  "transformers_version": "4.30.2"
+}

pytorch_model-00001-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b567345b37554b3e84c84587d3d66b97f1d4e2954399dae288cb9fe92db7e0e
+size 396364479

pytorch_model-00002-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:040b32d42253cfd2f94a000ea4563a12d9bcb2fe9b30b71ccb03d6b37ecd2b6d
+size 404770755

pytorch_model-00003-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c5f81d4a90d026761a95a9a95883e630bdaa6c4f3a3ae4fdb2c314d8da6e552
+size 404770755

pytorch_model-00004-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6469de41625275e6fbc42e2b7e1c13c62777694e28321d7d4bb2c5c0c668af30
+size 404770755

pytorch_model-00005-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44f87b5e389c534ab7687511f218afa794a0c1eb6cd10cce413fbdc3b82b53ad
+size 404770755

pytorch_model-00006-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef32773b02b2bac442b16dcc75ce8518816e9200330f5c4c9d2684a662e33a2c
+size 404770755

pytorch_model-00007-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5b3b34f6c7533cefc5eec6dfb41d27a15db68279c72322c9f2658a63f86761e
+size 404770755

pytorch_model-00008-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9742a8f3b8d673369c39c2f1b919ee38fb9a947ad046771c35bd04754b63e12d
+size 404770755

pytorch_model-00009-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ddfefe003828b77b667494671432b83258be6a5ad4e22e46f6e149cc2d755fd
+size 404770755

pytorch_model-00010-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d4d082b37ab603d1101a5483fbb3f89ab3ca9390bbcaa03b9d0b2fb47754100
+size 404770755

pytorch_model-00011-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1510f74784fae8ec105fb21d6ccdd965eb8651ea31801f7303e52f821f883820
+size 404770755

pytorch_model-00012-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:821dd9641581709bd132fcaa2f182950030a06431760f567134e12a6d26d407b
+size 404770755

pytorch_model-00013-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9a7f78aa05cf623ea67404b1dcfcce60e263c558fdaf5261d001f073480659c
+size 404770755

pytorch_model-00014-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c9d5742c113fcd492289fc487d072e7ccbc55662303165c0dbdb1ded84445d1
+size 404770755

pytorch_model-00015-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e37c1654eef3ba8381a9e23a5d737e4a87a9730fe627d92fc03470b5dde81ea4
+size 404770755

pytorch_model-00016-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f76680ba82e80bb4d301106f0f788c8250e67c4c30137d7ede867a56f558f138
+size 404770755

pytorch_model-00017-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52a8e15c1b40613e41b54933af1c1a4ecb454739d6f212cebb58a9b972eb9d95
+size 404770755

pytorch_model-00018-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c9b39cfdfc2c1a24f0c89b99f70058fc5bc0ee34ae2ff9ef12ccbc36341dd5e
+size 404770755

pytorch_model-00019-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c12e8c6480efef5ac094d3cc02bdca5d19796d931676f26776cbf2e678d6ad1
+size 404770755

pytorch_model-00020-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:087b16b7d96a97a4ca23c4e6ba213a46db5310a3b510112366d268ce0db4de9f
+size 404770755

pytorch_model-00021-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d01fc9179c646fd2105fa96a31ed81b906a2aa0eb8ab5f545ceffe7d3e459b9
+size 404770755

pytorch_model-00022-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b43e8290f4de9cd832a38955ebcfcedc15e88a6159ab94d6806f6442594aa5d9
+size 404770755

pytorch_model-00023-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef72145ac2dee9b13a376cf8b4bb73639b80662505b3fa65bb53fb0c9af4dd4c
+size 404770755

pytorch_model-00024-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a28ef5e9ca41670e6af69197e3bc1d13c14cfb7525f8d3eea8bb22b1d4b1055
+size 404770755

pytorch_model-00025-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcc1a666e55284023cde1dc4f2bcbade5ef13633590a398ef116c67103780d99
+size 404770755

pytorch_model-00026-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36e3dd91f25ee2668c8fa8ac55ce625a8a17a1a9aed380f8acc0651712362e4c
+size 404770755

pytorch_model-00027-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb803caeebceae2e4d83cfef61571253ae1edf5de48b4c8523a12a66ce3205d2
+size 404770755

pytorch_model-00028-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79271e20de9ec578c931a68f8e3455ff684a31acb6f82ed3eac5a2ca9821baff
+size 404770755

pytorch_model-00029-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33dfa6c15f0ca81a2e0d8b7e733790fd4dc164ad62440baf45e745da3a8d03d8
+size 404770755

pytorch_model-00030-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:919fc5b03034f78134a50cde7a0cb4493baf8c0f5ad7dcdeba5313059f852538
+size 404770755

pytorch_model-00031-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb6f00d3914045fccf9c04e5bf16796509af3641cf233457d1da531dab783a91
+size 404770755

pytorch_model-00032-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1ec1d12eb55e8cd66197aa46f09076ef1e0af09a1ab5d8743e4362bb59234c1
+size 404770755

pytorch_model-00033-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42d5976ecd4c3cd7500ffdad1d9660f46a4830b09739cefddae50db070153da0
+size 270559679

pytorch_model-00034-of-00034.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c379ef15678b8f07be84d699f9970e1be9cc99135b32c56420f1ab3363e3deee
+size 262144938

pytorch_model.bin.index.json

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