Upload with huggingface_hub

config.json

@@ -0,0 +1,23 @@
+{
+  "_name_or_path": "yahma/llama-7b-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": 2048,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "pad_token_id": 0,
+  "rms_norm_eps": 1e-06,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.28.1",
+  "use_cache": true,
+  "vocab_size": 32000
+}

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": "4.28.1"
+}

pytorch_model-00001-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:385c317fbb9063da8876a7beca94a9518662fd7c05db4b427e13132547c71870
+size 396364479

pytorch_model-00002-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d9b66f09375e6e6974dcf95ed15b7456a9332542e949407a8fdb0cd3ee36e70
+size 371215393

pytorch_model-00003-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a46aed19734340cd649e3e3fc03b862fb1f695619328b618be4a3f72befcef01
+size 371215986

pytorch_model-00004-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a49dd6bc23629732674d56d67abaf995b270a92b71074fbd14bb81261b4bafb8
+size 371215986

pytorch_model-00005-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee0c65efa7e16c73eb742a832800bd98f59d7cbdcd4919693f7c7d3494d30533
+size 371215986

pytorch_model-00006-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1350c790d02841971a008e919f6887b0af485760dd9727f3fe63313e6efb4df2
+size 314575888

pytorch_model-00007-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32fd211e91e2671d421b60454276fc287ffa1c922731fca465d7e10458e0b009
+size 314592882

pytorch_model-00008-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e1c13877446014d132cbf61bc5af8309901fb02e97ad5a068252ba874a2d18e
+size 314592882

pytorch_model-00009-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10ee2ace75822b4dd07cd8dec0c8760637f52951de7c53d9a753885fabd2be35
+size 371215393

pytorch_model-00010-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:684abc4aeb90ef6d7b009d3126fbcbbfe5ff4b76b100dfc03ee305315be57f13
+size 371215986

pytorch_model-00011-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82cfec476692463722fdaee83dc00bdb53e8f032f7de2660c5a0b668d09ce16a
+size 371215986

pytorch_model-00012-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74c6670aa32fdccc4c627d6d9b1b0e5e7814a46f778318d55ad418dd8b4940b3
+size 371215986

pytorch_model-00013-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec0629bc8781105f72dfa312a50076c0d56b824deea1ff8e3c99a99db8ecb02c
+size 314575888

pytorch_model-00014-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d82bd3a4c57e93ae5c26dc628dfa3f5045a911992c84ae049b6a7ec9d929a0c
+size 314592882

pytorch_model-00015-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e5888e005646b5be2ad02e5b1ec7bd3ed853a095308c15c3bd95b507421e6ca
+size 314592882

pytorch_model-00016-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bb78ed16bf91a496d35ac8699448a3a8a93d28fbfdacbf5074db4a78b22eb6e
+size 371215393

pytorch_model-00017-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:451ed62b0714ae8a200620a3346dd1430be3adeb254dec12bd5bc3f7ec1de755
+size 371215986

pytorch_model-00018-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4dd014b3f237d09894a778f1c929c4968cb114a85baf05cd3a2e4b2226f97bd
+size 371215986

pytorch_model-00019-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0de5d436a8447357939d4200c6d9ec278dcf13ab4f5417cd574dca983f1fefca
+size 371215986

pytorch_model-00020-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e696e6c6656374cf38a586a7b68bcbe3e12261434fc08dd985e4043fc7a9259f
+size 314575888

pytorch_model-00021-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0fc626d8a833be90fecec2b878448ef107fa52406fab468ba411c65617475db
+size 314592882

pytorch_model-00022-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87a1c72611d28716f13ba8dd56f2bcdf4c0d9a7b70a9cf2637e44f043ee6743c
+size 314592882

pytorch_model-00023-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15275a30e2651b790ee4cc1636fa0e9f9f0586c0ab26a17ba87998c18e066b46
+size 371215393

pytorch_model-00024-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15274ccd6500df8edb3ca84b3f594cb3b3f7480ddb60ecbc6e0119b138854e84
+size 371215986

pytorch_model-00025-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8381cc0ef25db2c9b1aae0d06edcb17156406e05f445537fc0209ddef9465887
+size 371215986

pytorch_model-00026-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0903a5813fbbc5cdaa7bc751dd46e4fc5f9766a00bf7b238a6927ed8b5d7e684
+size 371215986

pytorch_model-00027-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a9673c5ec0e7004ed81683e24078ad58a01f98745ae7c26b7380fce83657f27
+size 314575888

pytorch_model-00028-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d41152f872128c68323fe4cd591fbe4e0e53d6df6b7ff3c36562960e57bb71e
+size 314592882

pytorch_model-00029-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:719580704b0c93b4f89633a6f79b375490b2a1ec3e1c3ee0309a7d710ebd0cc5
+size 314592882

pytorch_model-00030-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e842357bbaaeae2241ab5e195b2aa6c4601588c5a4e5ca6cda2a227087d0eb3
+size 371215393

pytorch_model-00031-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68c47a766ef4cc7d9643586d043d8da57c8ec0bbfc3030ff3815adabe4ee4c94
+size 371215986

pytorch_model-00032-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c8139dbc24d45fe3f6a5fb5d26475f5a25bb7cd01623e04f01e373ef4ab73c0
+size 371215986

pytorch_model-00033-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c41dc95a84eafa17afc7cbbcb9e7e73a1557700577de81f622fdf17f64a70f72
+size 371215986

pytorch_model-00034-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00d0af9f6e0c7d8f4dada152eac95b6675e17b5f8bf39c3cb0f6a772daf17733
+size 314575888

pytorch_model-00035-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecb5c96784290070bf3a395155eb3f7a4cf78d4292b2d7bae30b44f9796bbad7
+size 314592882

pytorch_model-00036-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e9b4a6bed142ccaa48825dba677b9a766cc5f98eb7ca5a787b96d8e16e78fc9
+size 314592882

pytorch_model-00037-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ab72489461080d70f5c78e722a741f16725ca34c29553aa1aeaf46327ab5230
+size 371215393

pytorch_model-00038-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22c420de6bf2b4ebd738d9a4c65e3d2a8d8c7e36b68473cba9d6175f7f13cb2e
+size 304114977

pytorch_model-00039-of-00039.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3350fb71c2bc1e88d77d9685716fd2759e879137814a4120d98265a43f405919
+size 262144938

pytorch_model.bin.index.json

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