Upload LlamaForCausalLM

config.json

@@ -0,0 +1,23 @@
+{
+  "_name_or_path": "sharpbai/openchat",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 13824,
+  "max_position_embeddings": 2048,
+  "model_type": "llama",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 40,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.30.2",
+  "use_cache": true,
+  "vocab_size": 32001
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.30.2"
+}

pytorch_model-00001-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d5c49331bce857a31a1818b812e6ffcfbee0fa90c035d82a8c1661bfb1e3ec7
+size 537408191

pytorch_model-00002-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12a5a0810b5da0b6801d645afee740c1a67187664973ba4507cddf16e5d8bfc6
+size 634412995

pytorch_model-00003-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e15304ff4a1dd4ee244d7cdb4e0984c8145c5a48941a5c811d9a5b41f6a50c1
+size 634412995

pytorch_model-00004-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eed71393cd1d8c29ceb534e9cd1c4972c0b863e09c30ec9dfa2984d5f81c178a
+size 634412995

pytorch_model-00005-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48be6072305e2378ce1ba0ef331c225acfe3cd630ac5b96d4bb55ebe12c52f02
+size 634412995

pytorch_model-00006-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fef932116a5ea6201d41aa412a232e95f177eeadfa38f0ff514d3fd9f110e982
+size 634412995

pytorch_model-00007-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:730d52fbd9814afc2e0e4bc21fd533cd212b93bac0c22ce7669eba54cd06b7d9
+size 634412995

pytorch_model-00008-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba9c2c0c21774e33ded13f10d65599dbc61b14fa0557d8fa3cbe76d030addd31
+size 634412995

pytorch_model-00009-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3230f8dfc8bfdead6cbea4af622bc0f7003d5e7a829737d645c2dee94aff145f
+size 634412995

pytorch_model-00010-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c0736b4aa2e7de84caf83f9a333e486e50e221925872efb5db33a6fd5fabc93
+size 634412995

pytorch_model-00011-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a2f3547d7414fd9a89b88a7f9e020cc77c674caefeff93788d10730a753ec45
+size 634412995

pytorch_model-00012-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4917ba253b32514f86a2ab34241f693132be7cd65286a56a3db30d342f2aa74
+size 634412995

pytorch_model-00013-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b4c649da243d38d770f38ae6e45b8d5ed969652ab039d6d03bc661b6a075d1d
+size 634412995

pytorch_model-00014-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee6fb136f7657ec6d33199e1424555d8f3c2228e1869d41781d25f0d8dba5a28
+size 634412995

pytorch_model-00015-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43e560f10fb608bc62c8c9b332f137a92ec7e9304a23635411ffcaf9c304ce7f
+size 634412995

pytorch_model-00016-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db1d7ff4d51b8cd56e73153c7cec414786ec8c7175ad24a63f95cc0790cdaf35
+size 634412995

pytorch_model-00017-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a9e5f98ecb44d2431a53fd89dc2b9b339316dc358c446964339744eded11860
+size 634412995

pytorch_model-00018-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19cac122ba22b5cc114b7884666e4d5e9096e086ea86023c1e823996251ac805
+size 634412995

pytorch_model-00019-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8fe83a9f3ff297f368a209f99b0ae91fbbe0c39d47229c54906bf3eb6e6c477
+size 634412995

pytorch_model-00020-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e41ab20501af279bf920b968b5481850496d7526df325484a6a31f3bd07609aa
+size 634412995

pytorch_model-00021-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f834759275f3f72d3584e32f27015095e860e656ef600ac6a80dcf4d1fc432c9
+size 634412995

pytorch_model-00022-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d946ebc734b6f0aef42e2ff48cce9c9efed591ad6321decb6623739b21daa39
+size 634412995

pytorch_model-00023-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87ad745cc19a8ceab24e856a984bd741d052e820af9f6c4c5b0abb0d1aa7b384
+size 634412995

pytorch_model-00024-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ee8c2a4b36d398417e091e9590f62380ce2a1e814aff13b0c4c061c5927f044
+size 634412995

pytorch_model-00025-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:295f0c7920395583645d9ee9240d14b543232a3e7d39c6b750e652bdea810007
+size 634412995

pytorch_model-00026-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33b28d0f9d6bfc1801858e937def27aac9c2742259f1cedf21e4b01ba76fa6b5
+size 634412995

pytorch_model-00027-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:314016801a4a89925aa7156c8e7a77e2b118697ebb847815cfff9549aade9c17
+size 634412995

pytorch_model-00028-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65a1113ecaa2d0de29b59bfb8542db47db447fb946c74fd8ed1841a3530f4989
+size 634412995

pytorch_model-00029-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e31b6ef62fa680e607a178f04deca9e4152da1a08048a19f2a5b67cf45f7f30b
+size 634412995

pytorch_model-00030-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14952ab26c50facfe2154e14e534cb8a5c668ae015bcced516f74fe6e5fc0f99
+size 634412995

pytorch_model-00031-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d20702cbd03e23ea94d18b44b64997a7382ae91fa5c7fd3af3789eb41c24249
+size 634412995

pytorch_model-00032-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88dbf47fa015fc97e513ee2b2b301c2f9331e7e9012b0892439e6b7bf106637a
+size 634412995

pytorch_model-00033-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7a820e01614c720f99a4ca1f4ff6c7746b29a042bd99190ef18cb01bdcf47e2
+size 634412995

pytorch_model-00034-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a78b2ef262da6e62294c0adbc6890b4f09f85938e52c404dc8c24a96a5fdc51
+size 634412995

pytorch_model-00035-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:354c7a345e9c1fa1eee683c2370688d55485c6dc473f6043771843f9dd3086ef
+size 634412995

pytorch_model-00036-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6663a855e82318f664bb103435e7606620e0c391479db952973bda69efc0b76f
+size 634412995

pytorch_model-00037-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5768b59122cdf8e0c4b94460790bc4a1333c06d0942e85e95964a68e605ae59d
+size 634412995

pytorch_model-00038-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d5ab2d1c01c58573e7f4ff679bee634eb1cdfa3ee8590a6ef82e2972ce01a5d
+size 634412995

pytorch_model-00039-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c575322871d735faf417ac6648c52209db019fbd6f7f1ae24a1edb558da0def
+size 634412995

pytorch_model-00040-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a60e200e69fdcc078aa85cd0fdbba1444913a6a0f977d3d27355b02e4a99c7e0
+size 634412995

pytorch_model-00041-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ff607dd09a9024d212bf2be61cdcb414be63024c71b6699ae2a06d2f2c11dc3
+size 424706495

pytorch_model-00042-of-00042.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdc656a849c8907779b36bb08a527efac58b94d81ebdd85b0e20744474c2f491
+size 327691178

pytorch_model.bin.index.json

@@ -0,0 +1,410 @@
+{
+  "metadata": {
+    "total_size": 26031759360
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00042-of-00042.bin",
+    "model.embed_tokens.weight": "pytorch_model-00001-of-00042.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00042.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00042.bin",
+    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00042.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00042.bin",
+    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00042.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00042.bin",
+    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00002-of-00042.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00011-of-00042.bin",
+    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00012-of-00042.bin",
+    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00012-of-00042.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00013-of-00042.bin",
+    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00013-of-00042.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00014-of-00042.bin",
+    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00014-of-00042.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00015-of-00042.bin",
+    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00015-of-00042.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00016-of-00042.bin",
+    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00016-of-00042.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00017-of-00042.bin",
+    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00017-of-00042.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00018-of-00042.bin",
+    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00018-of-00042.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00019-of-00042.bin",
+    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00019-of-00042.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00020-of-00042.bin",
+    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00020-of-00042.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00042.bin",
+    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00003-of-00042.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00021-of-00042.bin",
+    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00021-of-00042.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00022-of-00042.bin",
+    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00022-of-00042.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00023-of-00042.bin",
+    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00023-of-00042.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00024-of-00042.bin",
+    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00024-of-00042.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00025-of-00042.bin",
+    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00025-of-00042.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00026-of-00042.bin",
+    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00026-of-00042.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00027-of-00042.bin",
+    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00027-of-00042.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00028-of-00042.bin",
+    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00028-of-00042.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00029-of-00042.bin",
+    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00029-of-00042.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00030-of-00042.bin",
+    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00030-of-00042.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00004-of-00042.bin",
+    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00004-of-00042.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00031-of-00042.bin",
+    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00031-of-00042.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00032-of-00042.bin",
+    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00032-of-00042.bin",
+    "model.layers.32.input_layernorm.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.32.mlp.down_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.32.mlp.gate_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.32.mlp.up_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.32.post_attention_layernorm.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.32.self_attn.k_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.32.self_attn.o_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.32.self_attn.q_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.32.self_attn.rotary_emb.inv_freq": "pytorch_model-00033-of-00042.bin",
+    "model.layers.32.self_attn.v_proj.weight": "pytorch_model-00033-of-00042.bin",
+    "model.layers.33.input_layernorm.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.33.mlp.down_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.33.mlp.gate_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.33.mlp.up_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.33.post_attention_layernorm.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.33.self_attn.k_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.33.self_attn.o_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.33.self_attn.q_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.33.self_attn.rotary_emb.inv_freq": "pytorch_model-00034-of-00042.bin",
+    "model.layers.33.self_attn.v_proj.weight": "pytorch_model-00034-of-00042.bin",
+    "model.layers.34.input_layernorm.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.34.mlp.down_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.34.mlp.gate_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.34.mlp.up_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.34.post_attention_layernorm.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.34.self_attn.k_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.34.self_attn.o_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.34.self_attn.q_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.34.self_attn.rotary_emb.inv_freq": "pytorch_model-00035-of-00042.bin",
+    "model.layers.34.self_attn.v_proj.weight": "pytorch_model-00035-of-00042.bin",
+    "model.layers.35.input_layernorm.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.35.mlp.down_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.35.mlp.gate_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.35.mlp.up_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.35.post_attention_layernorm.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.35.self_attn.k_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.35.self_attn.o_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.35.self_attn.q_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.35.self_attn.rotary_emb.inv_freq": "pytorch_model-00036-of-00042.bin",
+    "model.layers.35.self_attn.v_proj.weight": "pytorch_model-00036-of-00042.bin",
+    "model.layers.36.input_layernorm.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.36.mlp.down_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.36.mlp.gate_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.36.mlp.up_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.36.post_attention_layernorm.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.36.self_attn.k_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.36.self_attn.o_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.36.self_attn.q_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.36.self_attn.rotary_emb.inv_freq": "pytorch_model-00037-of-00042.bin",
+    "model.layers.36.self_attn.v_proj.weight": "pytorch_model-00037-of-00042.bin",
+    "model.layers.37.input_layernorm.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.37.mlp.down_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.37.mlp.gate_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.37.mlp.up_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.37.post_attention_layernorm.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.37.self_attn.k_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.37.self_attn.o_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.37.self_attn.q_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.37.self_attn.rotary_emb.inv_freq": "pytorch_model-00038-of-00042.bin",
+    "model.layers.37.self_attn.v_proj.weight": "pytorch_model-00038-of-00042.bin",
+    "model.layers.38.input_layernorm.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.38.mlp.down_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.38.mlp.gate_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.38.mlp.up_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.38.post_attention_layernorm.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.38.self_attn.k_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.38.self_attn.o_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.38.self_attn.q_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.38.self_attn.rotary_emb.inv_freq": "pytorch_model-00039-of-00042.bin",
+    "model.layers.38.self_attn.v_proj.weight": "pytorch_model-00039-of-00042.bin",
+    "model.layers.39.input_layernorm.weight": "pytorch_model-00041-of-00042.bin",
+    "model.layers.39.mlp.down_proj.weight": "pytorch_model-00041-of-00042.bin",
+    "model.layers.39.mlp.gate_proj.weight": "pytorch_model-00041-of-00042.bin",
+    "model.layers.39.mlp.up_proj.weight": "pytorch_model-00041-of-00042.bin",
+    "model.layers.39.post_attention_layernorm.weight": "pytorch_model-00041-of-00042.bin",
+    "model.layers.39.self_attn.k_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.39.self_attn.o_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.39.self_attn.q_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.39.self_attn.rotary_emb.inv_freq": "pytorch_model-00040-of-00042.bin",
+    "model.layers.39.self_attn.v_proj.weight": "pytorch_model-00040-of-00042.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00005-of-00042.bin",
+    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00005-of-00042.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00006-of-00042.bin",
+    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00006-of-00042.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00007-of-00042.bin",
+    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00007-of-00042.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00008-of-00042.bin",
+    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00008-of-00042.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00009-of-00042.bin",
+    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00009-of-00042.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00011-of-00042.bin",
+    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00010-of-00042.bin",
+    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00010-of-00042.bin",
+    "model.norm.weight": "pytorch_model-00041-of-00042.bin"
+  }
+}