Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

config.json

@@ -0,0 +1,30 @@
+{
+  "architectures": [
+    "Qwen3ForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 8192,
+  "initializer_range": 0.02,
+  "intermediate_size": 29568,
+  "max_position_embeddings": 40960,
+  "max_window_layers": 64,
+  "model_type": "qwen3",
+  "num_attention_heads": 64,
+  "num_hidden_layers": 80,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "vocab_size": 151936
+}

model-00001-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61d8531d1f81d1383acf3bbbac9cc86799c1ab723f4ca6d6e3b0f9a9b9cbbd25
+size 4978672024

model-00002-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d6b57a4d4e5f727556f8f0dd5c5dcbf45746d450265b09a909ea1fd4d9b1180
+size 4980822416

model-00003-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e449ae2d61491564ab3f88a2a07a2680766d322ae233dc9b8bcbe908c9056f60
+size 4582381232

model-00004-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48fbe2ced8e5d7f3f84b8be1f2a64986151c637c48400617db7ae615adb4975b
+size 4964044712

model-00005-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8592a3b57d0a67b8d8fce74f6b84cecc5096f250104fe5dba621cbd149389d03
+size 4781593584

model-00006-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4de35e7d8a69581677277a1fe318635d03ebe1ec92fac250f81ccf45c44832a6
+size 4980838936

model-00007-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d63efa2cefba168ce668100cbe8c3b0569d339d794f6c59df0032f79d8656df
+size 4582381264

model-00008-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:204c31141374e12802b99f554238889608219c16dca1cecbd946c32ce2ff6f9f
+size 4964028240

model-00009-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93794680f27d50d028a1c99cff2eae192aff3e61466022e254e7042caddbdc2d
+size 4781626592

model-00010-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fdf85fbbcbec6581335017ecb8c4c1d4b047b71d6c2c385f88281f980a78f04
+size 4980822448

model-00011-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4cd0a5d9651fc545b144793285ab67899a4e0143f626053e83dd6df7cdaf4ad8
+size 4764815896

model-00012-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c0ed72485fefc060baf3e4aec6e4a201ae6f52409740dd71baba212d6bb03c73
+size 4781610096

model-00013-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:045224aa955cc8ba6785401226d2e816d3d7cbb26647be51cc4ec7bcc702bea7
+size 4781610104

model-00014-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d68bfa4d4d99c27033d9b028f83cad1ea32c80ee6ea21a33569b3a0aa8c3695e
+size 4980822448

model-00015-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad7858bab59b3e4f76bc5cf046656de51899e9a9416dda5cf2bd0c4a0ab52fbc
+size 4764815896

model-00016-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:350e0b89e9a0f93f9e1f7ba1e6f830c41c38ffdf813bb10b15f9ecdf810b8258
+size 4781610096

model-00017-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:503fa67d7d5bb56599f46fb7ad4a5e7e3d8f0a79d9d02a51e608e50fe71443d8
+size 4781610104

model-00018-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4945e879dc16e52f53cd72850c88bad8f0eb7627ca7e0cab6238bd674a1e0a7e
+size 4980822448

model-00019-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a276ac969178765a11b24009bb26a4d794c46dc0366957e76ca66a5e0d026bd5
+size 4764815896

model-00020-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19b841082f9bfcf23a1beed49631e60d90e3dabcd2ffafbcf2653f0eddc9d313
+size 4932606008

model-00021-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b780c1c411b0b153da5a06313b17145373ccbd129548cc37c974e5582f51c4a
+size 4630614192

model-00022-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cce4c30cc5efad79a98e7a0eee32fd6b9425d3181ab1fc986dea546207db4785
+size 4781593616

model-00023-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99c19833f5be9d4a3640dbb4da8b4145f465b1ac3e5417e3fd0828b13ce59c2b
+size 4964044736

model-00024-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e571dacedd834da1cc739297b61a7adaa2a9e8edc26b6d76c3c5d6aa52017d9
+size 4781610096

model-00025-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cb4457897c90b6c30fb1687107f13c8b1b9b6bccafb84478f00f9a4891378ad
+size 4781610104

model-00026-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc6efe9bd5d36485f6e6ed48cec49739209d644571d862979e1a07707cdfd3e6
+size 4980822448

model-00027-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d0d0da582857f590f8b4c34cb176a6ddc6cf80fc55125da7ff249fa8234ccd7
+size 4764815896

model-00028-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4cf90c45bb46f9c2c29adadf00acdb41ff760309296025e12930f885c882dc6
+size 4781610096

model-00029-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b96092f2efa70a7ece0b01f4d2286fc6f3583a311244d2f9b823003db62384d
+size 4781610104

model-00030-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:352997a6410ad74fcab769182d1a8de0c248d2d79fc3be3dd366b46de4748528
+size 4781593616

model-00031-of-00031.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e52be5d002bb99434fbb60481509eafcfe548966f2f73b6fc344713c955b82b6
+size 484442232

model.safetensors.index.json

@@ -0,0 +1,890 @@
+{
+  "metadata": {
+    "total_size": 145406615552
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00001-of-00031.safetensors",
+    "model.norm.weight": "model-00001-of-00031.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00031.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.k_norm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.q_norm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.k_norm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.q_norm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.k_norm.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00002-of-00031.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.2.self_attn.q_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.k_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.q_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.k_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.q_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.k_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.q_norm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00003-of-00031.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.k_norm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.q_norm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.k_norm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.q_norm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00004-of-00031.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.k_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.q_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.k_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.q_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.k_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.q_norm.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00005-of-00031.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.k_norm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.q_norm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.k_norm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.q_norm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.k_norm.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00006-of-00031.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.13.self_attn.q_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.k_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.q_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.k_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.q_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.k_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.q_norm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00007-of-00031.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.k_norm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.q_norm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.k_norm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.q_norm.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00008-of-00031.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.k_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.q_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.k_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.q_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.k_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.q_norm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00009-of-00031.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.k_norm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.q_norm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.k_norm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.q_norm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.k_norm.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00010-of-00031.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.24.self_attn.q_norm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.k_norm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.q_norm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.k_norm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.q_norm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00011-of-00031.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.k_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.q_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.k_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.q_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.k_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.q_norm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00012-of-00031.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.k_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.q_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.k_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.q_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.k_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.q_norm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-00013-of-00031.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.k_norm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.q_norm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.k_norm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.q_norm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.k_norm.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-00014-of-00031.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.35.self_attn.q_norm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.k_norm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.q_norm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.k_norm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.q_norm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-00015-of-00031.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.k_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.q_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.k_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.q_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.k_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.q_norm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-00016-of-00031.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.k_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.q_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.mlp.down_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.mlp.gate_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.k_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.q_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.42.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.down_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.gate_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.mlp.up_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.post_attention_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.k_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.k_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.o_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.q_norm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.q_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.self_attn.v_proj.weight": "model-00017-of-00031.safetensors",
+    "model.layers.43.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.input_layernorm.weight": "model-00017-of-00031.safetensors",
+    "model.layers.44.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.k_norm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.k_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.o_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.q_norm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.q_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.44.self_attn.v_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.input_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.k_norm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.k_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.o_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.q_norm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.q_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.45.self_attn.v_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.input_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.down_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.gate_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.mlp.up_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.post_attention_layernorm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.k_norm.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.k_proj.weight": "model-00018-of-00031.safetensors",
+    "model.layers.46.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.46.self_attn.q_norm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.46.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.46.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.input_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.mlp.down_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.mlp.up_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.post_attention_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.k_norm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.k_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.q_norm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.47.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.input_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.down_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.mlp.up_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.post_attention_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.k_norm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.k_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.o_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.q_norm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.q_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.48.self_attn.v_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.input_layernorm.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.down_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.gate_proj.weight": "model-00019-of-00031.safetensors",
+    "model.layers.49.mlp.up_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.k_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.q_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.49.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.mlp.down_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.mlp.up_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.k_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.q_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.50.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.down_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.mlp.up_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.post_attention_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.k_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.q_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.q_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.51.self_attn.v_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.input_layernorm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.mlp.gate_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.k_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.k_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.o_proj.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.q_norm.weight": "model-00020-of-00031.safetensors",
+    "model.layers.52.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.52.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.52.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.52.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.52.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.mlp.gate_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.k_norm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.k_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.o_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.q_norm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.53.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.down_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.gate_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.mlp.up_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.post_attention_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.k_norm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.k_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.o_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.q_norm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.q_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.54.self_attn.v_proj.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.input_layernorm.weight": "model-00021-of-00031.safetensors",
+    "model.layers.55.mlp.down_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.k_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.k_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.o_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.q_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.q_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.55.self_attn.v_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.input_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.down_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.k_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.k_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.o_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.q_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.q_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.56.self_attn.v_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.gate_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.up_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.post_attention_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.k_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.k_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.o_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.q_norm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.q_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.self_attn.v_proj.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.input_layernorm.weight": "model-00022-of-00031.safetensors",
+    "model.layers.57.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.input_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.mlp.up_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.post_attention_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.k_norm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.k_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.o_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.q_norm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.q_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.58.self_attn.v_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.input_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.mlp.up_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.post_attention_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.k_norm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.k_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.o_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.q_norm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.q_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.59.self_attn.v_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.input_layernorm.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.down_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.gate_proj.weight": "model-00023-of-00031.safetensors",
+    "model.layers.60.mlp.up_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.k_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.q_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.60.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.mlp.gate_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.mlp.up_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.k_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.q_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.61.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.o_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.q_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.q_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.v_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.gate_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.mlp.up_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.post_attention_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.k_norm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.62.self_attn.k_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.input_layernorm.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.mlp.down_proj.weight": "model-00024-of-00031.safetensors",
+    "model.layers.63.mlp.gate_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.k_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.q_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.63.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.mlp.down_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.mlp.gate_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.k_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.q_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.64.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.down_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.gate_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.mlp.up_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.post_attention_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.k_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.k_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.o_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.q_norm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.q_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.65.self_attn.v_proj.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.input_layernorm.weight": "model-00025-of-00031.safetensors",
+    "model.layers.66.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.k_norm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.k_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.o_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.q_norm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.q_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.66.self_attn.v_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.input_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.k_norm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.k_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.o_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.q_norm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.q_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.67.self_attn.v_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.input_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.down_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.gate_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.mlp.up_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.post_attention_layernorm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.k_norm.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.k_proj.weight": "model-00026-of-00031.safetensors",
+    "model.layers.68.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.68.self_attn.q_norm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.68.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.68.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.input_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.mlp.down_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.mlp.up_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.post_attention_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.k_norm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.k_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.q_norm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.69.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.input_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.down_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.mlp.up_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.post_attention_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.k_norm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.k_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.o_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.q_norm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.q_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.70.self_attn.v_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.input_layernorm.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.down_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.gate_proj.weight": "model-00027-of-00031.safetensors",
+    "model.layers.71.mlp.up_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.k_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.q_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.71.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.mlp.gate_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.mlp.up_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.k_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.q_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.72.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.o_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.q_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.q_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.v_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.gate_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.mlp.up_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.post_attention_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.k_norm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.73.self_attn.k_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.input_layernorm.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.mlp.down_proj.weight": "model-00028-of-00031.safetensors",
+    "model.layers.74.mlp.gate_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.k_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.q_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.74.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.mlp.down_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.mlp.gate_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.k_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.q_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.75.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.down_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.gate_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.mlp.up_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.post_attention_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.k_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.k_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.o_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.q_norm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.q_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.76.self_attn.v_proj.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.input_layernorm.weight": "model-00029-of-00031.safetensors",
+    "model.layers.77.mlp.down_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.mlp.gate_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.mlp.up_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.post_attention_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.k_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.k_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.o_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.q_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.q_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.77.self_attn.v_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.input_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.down_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.gate_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.mlp.up_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.post_attention_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.k_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.k_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.o_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.q_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.q_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.78.self_attn.v_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.gate_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.up_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.post_attention_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.k_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.k_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.o_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.q_norm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.q_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.self_attn.v_proj.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.input_layernorm.weight": "model-00030-of-00031.safetensors",
+    "model.layers.79.mlp.down_proj.weight": "model-00031-of-00031.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

stage2_metadata.json

@@ -0,0 +1,32 @@
+{
+  "stage": "2-duplicate",
+  "source_model": "./Qwen3-32B-to-72B-Stage1",
+  "method": "Simple layer duplication",
+  "layer_mapping": {
+    "0-23": "0-23 (unchanged)",
+    "24-39": "24-55 (each duplicated once)",
+    "40-63": "56-79 (unchanged)"
+  },
+  "duplication_info": {
+    "method": "exact_copy",
+    "layers_duplicated": [
+      24,
+      25,
+      26,
+      27,
+      28,
+      29,
+      30,
+      31,
+      32,
+      33,
+      34,
+      35,
+      36,
+      37,
+      38,
+      39
+    ]
+  },
+  "final_layers": 80
+}

stage2_v3.py

@@ -0,0 +1,568 @@
+#!/usr/bin/env python
+"""
+Stage 2: Expand Qwen3 from 64 to 80 layers using simple duplication
+Mapping:
+- Layers 0-23 → 0-23 (unchanged)
+- Layers 24-39 → 24-55 (each layer duplicated once)
+- Layers 40-63 → 56-79 (unchanged)
+"""
+
+import torch
+import os
+import json
+from tqdm import tqdm
+from transformers import AutoConfig, AutoModelForCausalLM, AutoTokenizer
+from safetensors.torch import load_file, save_file
+import numpy as np
+from collections import OrderedDict
+import gc
+import shutil
+
+# Configuration
+INPUT_DIR = "./Qwen3-32B-to-72B-Stage1"  # Output from stage 1
+OUTPUT_DIR = "./Qwen3-72B-DupeLayers"
+TARGET_LAYERS = 80
+SOURCE_LAYERS = 64
+
+def load_model_sharted(model_path):
+    """Load model weights from sharted safetensors files."""
+    print("\n💩 Loading sharted weights...")
+    
+    index_path = os.path.join(model_path, "model.safetensors.index.json")
+    
+    if not os.path.exists(index_path):
+        raise FileNotFoundError(f"No index file found at {index_path}")
+    
+    with open(index_path, 'r') as f:
+        index = json.load(f)
+    
+    weight_map = index['weight_map']
+    unique_files = set(weight_map.values())
+    
+    all_weights = {}
+    for file in tqdm(unique_files, desc="Loading sharts"):
+        file_path = os.path.join(model_path, file)
+        weights = load_file(file_path)
+        all_weights.update(weights)
+    
+    return all_weights
+
+def save_model_sharted(state_dict, output_dir, max_shart_size="5GB"):
+    """Save model in sharted safetensors format."""
+    print("\n💩 Sharting model weights...")
+    
+    os.makedirs(output_dir, exist_ok=True)
+    
+    # Convert max_shart_size to bytes
+    size_map = {'GB': 1e9, 'MB': 1e6}
+    for unit, multiplier in size_map.items():
+        if unit in max_shart_size:
+            max_bytes = int(float(max_shart_size.replace(unit, '')) * multiplier)
+            break
+    
+    # Group weights into sharts
+    sharts = []
+    current_shart = {}
+    current_size = 0
+    
+    for name, tensor in state_dict.items():
+        tensor_size = tensor.numel() * tensor.element_size()
+        
+        if current_size + tensor_size > max_bytes and current_shart:
+            sharts.append(current_shart)
+            current_shart = {}
+            current_size = 0
+        
+        current_shart[name] = tensor
+        current_size += tensor_size
+    
+    if current_shart:
+        sharts.append(current_shart)
+    
+    # Save sharts
+    weight_map = {}
+    for i, shart in enumerate(tqdm(sharts, desc="Saving sharts")):
+        shart_name = f"model-{i+1:05d}-of-{len(sharts):05d}.safetensors"
+        save_file(shart, os.path.join(output_dir, shart_name))
+        
+        for name in shart:
+            weight_map[name] = shart_name
+    
+    # Save index
+    index = {
+        "metadata": {"total_size": sum(t.numel() * t.element_size() for t in state_dict.values())},
+        "weight_map": weight_map
+    }
+    
+    with open(os.path.join(output_dir, "model.safetensors.index.json"), 'w') as f:
+        json.dump(index, f, indent=2)
+    
+    print(f"💩 Successfully sharted into {len(sharts)} files!")
+
+def extract_layer_weights(weights, layer_idx):
+    """Extract all weights for a specific layer."""
+    layer_weights = OrderedDict()
+    prefix = f"model.layers.{layer_idx}."
+    
+    for name, tensor in weights.items():
+        if name.startswith(prefix):
+            # Remove the layer prefix to get the component name
+            component_name = name[len(prefix):]
+            layer_weights[component_name] = tensor
+    
+    return layer_weights
+
+def create_layer_weights(layer_weights, new_layer_idx):
+    """Create weight dict with new layer index."""
+    result = OrderedDict()
+    prefix = f"model.layers.{new_layer_idx}."
+    
+    for component_name, tensor in layer_weights.items():
+        full_name = prefix + component_name
+        result[full_name] = tensor.clone()  # Clone to ensure independent copies
+    
+    return result
+
+def verify_architecture(model_path):
+    """Verify the model architecture matches expected Qwen3-72B dimensions."""
+    print("\n" + "="*60)
+    print("ARCHITECTURE VERIFICATION")
+    print("="*60)
+    
+    print("\nLoading model for verification...")
+    model = AutoModelForCausalLM.from_pretrained(
+        model_path,
+        torch_dtype=torch.bfloat16,
+        device_map="cpu",
+        trust_remote_code=True
+    )
+    
+    expected = {
+        "lm_head.weight": (151936, 8192),
+        "model.embed_tokens.weight": (151936, 8192),
+        "model.layers.0.input_layernorm.weight": (8192,),
+        "model.layers.0.mlp.down_proj.weight": (8192, 29568),
+        "model.layers.0.mlp.gate_proj.weight": (29568, 8192),
+        "model.layers.0.mlp.up_proj.weight": (29568, 8192),
+        "model.layers.0.post_attention_layernorm.weight": (8192,),
+        "model.layers.0.self_attn.k_norm.weight": (128,),
+        "model.layers.0.self_attn.k_proj.weight": (1024, 8192),
+        "model.layers.0.self_attn.o_proj.weight": (8192, 8192),
+        "model.layers.0.self_attn.q_norm.weight": (128,),
+        "model.layers.0.self_attn.q_proj.weight": (8192, 8192),
+        "model.layers.0.self_attn.v_proj.weight": (1024, 8192),
+        "model.norm.weight": (8192,),
+    }
+    
+    all_correct = True
+    
+    # Check specific layers including duplicated ones
+    check_layers = [0, 24, 25, 39, 40, 56, 79]  # Original and duplicated layers
+    
+    for layer_idx in check_layers:
+        print(f"\n📍 Checking layer {layer_idx}:")
+        for base_name, expected_shape in expected.items():
+            if "layers.0." in base_name:
+                name = base_name.replace("layers.0.", f"layers.{layer_idx}.")
+                param_dict = dict(model.named_parameters())
+                if name in param_dict:
+                    actual_shape = tuple(param_dict[name].shape)
+                    if actual_shape == expected_shape:
+                        print(f"   ✓ {name.split('.')[-1]}: {actual_shape}")
+                    else:
+                        print(f"   ✗ {name}: {actual_shape} (expected {expected_shape})")
+                        all_correct = False
+    
+    num_layers = model.config.num_hidden_layers
+    print(f"\nTotal layers: {num_layers} (expected: 80)")
+    
+    if all_correct and num_layers == 80:
+        print("\n✅ Architecture verification PASSED!")
+    else:
+        print("\n❌ Architecture verification FAILED!")
+    
+    del model
+    torch.cuda.empty_cache()
+    return all_correct
+
+def run_diagnostics(model_path):
+    """Run comprehensive diagnostics on the expanded model."""
+    print("\n" + "="*60)
+    print("COMPREHENSIVE DIAGNOSTICS")
+    print("="*60)
+    
+    # Load model and tokenizer
+    print("\nLoading model for diagnostics...")
+    model = AutoModelForCausalLM.from_pretrained(
+        model_path,
+        torch_dtype=torch.bfloat16,
+        device_map="auto",
+        trust_remote_code=True
+    )
+    tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
+    
+    # Test generation quality
+    print("\n🧪 Generation Quality Tests:")
+    test_cases = [
+        ("The capital of France is", ["Paris"]),
+        ("2 + 2 =", ["4", "four"]),
+        ("The quick brown fox", ["jumps", "jumped", "lazy", "dog"]),
+        ("Hello, my name is", None),
+        ("Water boils at", ["100", "212", "degrees"]),
+        ("The Earth orbits the", ["Sun", "solar"]),
+        ("Machine learning is a type of", ["artificial intelligence", "AI"]),
+        ("Python is a", ["programming", "language", "snake"]),
+        ("The largest planet is", ["Jupiter"]),
+        ("DNA stands for", ["deoxyribonucleic", "acid"]),
+        # Additional tests
+        ("The derivative of x squared is", ["2x", "two"]),
+        ("Shakespeare wrote", ["plays", "Hamlet", "Romeo"]),
+        ("The speed of light is", ["299", "300", "fast"]),
+        ("Photosynthesis converts", ["light", "energy", "carbon"]),
+        ("The Pythagorean theorem states", ["a²", "squared", "hypotenuse"]),
+    ]
+    
+    device = model.device
+    coherent_count = 0
+    total_tests = len(test_cases)
+    
+    for prompt, expected in test_cases:
+        inputs = tokenizer(prompt, return_tensors="pt").to(device)
+        
+        with torch.no_grad():
+            outputs = model.generate(
+                **inputs,
+                max_new_tokens=20,
+                do_sample=True,
+                temperature=0.7,
+                top_k=50,
+                top_p=0.95,
+                pad_token_id=tokenizer.pad_token_id,
+            )
+        
+        generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
+        generated_only = generated_text[len(prompt):].strip()
+        
+        print(f"\n   Prompt: '{prompt}'")
+        print(f"   Generated: '{generated_only}'")
+        
+        # Check coherence
+        is_coherent = True
+        
+        # Check for repetition
+        words = generated_only.split()
+        if len(words) > 3:
+            if len(set(words)) < len(words) / 2:
+                print("   ⚠️  High repetition detected")
+                is_coherent = False
+        
+        # Check for expected content
+        if expected and len(generated_only) > 0:
+            found = any(kw.lower() in generated_only.lower() for kw in expected)
+            if found:
+                print("   ✓ Contains expected content")
+            else:
+                print("   ⚠️  Missing expected keywords")
+                is_coherent = False
+        
+        if is_coherent and len(generated_only.split()) >= 2:
+            coherent_count += 1
+    
+    coherence_rate = (coherent_count / total_tests) * 100
+    print(f"\n📊 Overall coherence rate: {coherence_rate:.1f}%")
+    
+    # Perplexity test
+    print("\n📈 Perplexity Test:")
+    test_texts = [
+        "The quick brown fox jumps over the lazy dog.",
+        "In the beginning was the Word, and the Word was with God.",
+        "To be or not to be, that is the question.",
+        "E equals m c squared is Einstein's famous equation.",
+    ]
+    
+    perplexities = []
+    for test_text in test_texts:
+        inputs = tokenizer(test_text, return_tensors="pt").to(device)
+        
+        with torch.no_grad():
+            outputs = model(**inputs, labels=inputs["input_ids"])
+            perplexity = torch.exp(outputs.loss).item()
+            perplexities.append(perplexity)
+        
+        print(f"   '{test_text[:30]}...': {perplexity:.2f}")
+    
+    avg_perplexity = np.mean(perplexities)
+    print(f"\n   Average perplexity: {avg_perplexity:.2f}")
+    
+    if avg_perplexity > 100:
+        print("   ⚠️  Very high perplexity")
+    elif avg_perplexity > 50:
+        print("   ⚠️  Moderately high perplexity")
+    else:
+        print("   ✓ Reasonable perplexity")
+    
+    # Test duplicate layer behavior
+    print("\n🔬 Duplicate Layer Analysis:")
+    print("Checking if duplicated layers maintain reasonable behavior...")
+    
+    # Get activations from a few layers
+    test_input = "The meaning of life is"
+    inputs = tokenizer(test_input, return_tensors="pt").to(device)
+    
+    activations = {}
+    hooks = []
+    
+    def get_activation(name):
+        def hook(model, input, output):
+            activations[name] = output[0].detach()
+        return hook
+    
+    # Register hooks for duplicate pairs
+    for layer_idx in [24, 25, 39, 40]:  # Original and duplicate
+        hook = model.model.layers[layer_idx].register_forward_hook(
+            get_activation(f'layer_{layer_idx}')
+        )
+        hooks.append(hook)
+    
+    with torch.no_grad():
+        _ = model(**inputs)
+    
+    # Remove hooks
+    for hook in hooks:
+        hook.remove()
+    
+    # Check similarity of duplicates
+    if len(activations) >= 4:
+        # Check 24 vs 25 (should be duplicates)
+        act_24 = activations['layer_24'].flatten()
+        act_25 = activations['layer_25'].flatten()
+        similarity_24_25 = torch.cosine_similarity(act_24.unsqueeze(0), act_25.unsqueeze(0)).item()
+        
+        # Check 39 vs 40 (should be different - 40 is original layer 40, not duplicate)
+        act_39 = activations['layer_39'].flatten()
+        act_40 = activations['layer_40'].flatten()
+        similarity_39_40 = torch.cosine_similarity(act_39.unsqueeze(0), act_40.unsqueeze(0)).item()
+        
+        print(f"   Cosine similarity layer 24 vs 25 (duplicate): {similarity_24_25:.4f}")
+        print(f"   Cosine similarity layer 39 vs 40 (different): {similarity_39_40:.4f}")
+        
+        if similarity_24_25 > 0.95:
+            print("   ✓ Duplicate layers show expected high similarity")
+        else:
+            print("   ⚠️  Duplicate layers diverged more than expected")
+    
+    # Weight statistics check
+    print("\n🔍 Weight Statistics (checking for anomalies):")
+    anomalies = 0
+    
+    for name, param in model.named_parameters():
+        if torch.isnan(param).any():
+            print(f"   ⚠️  {name}: Contains NaN!")
+            anomalies += 1
+        elif torch.isinf(param).any():
+            print(f"   ⚠️  {name}: Contains Inf!")
+            anomalies += 1
+        elif param.std() < 1e-8:
+            print(f"   ⚠️  {name}: Zero variance!")
+            anomalies += 1
+    
+    if anomalies == 0:
+        print("   ✓ No anomalies detected in weights")
+    
+    # Final summary
+    success = coherence_rate >= 60 and avg_perplexity < 100 and anomalies == 0
+    
+    print("\n" + "="*60)
+    print("DIAGNOSTIC SUMMARY")
+    print("="*60)
+    
+    if success:
+        print("✅ Model passed all diagnostics!")
+        print("   - Good coherence rate")
+        print("   - Reasonable perplexity")
+        print("   - No weight anomalies")
+        print("   - Duplicate layers functioning correctly")
+    else:
+        print("⚠️  Some issues detected:")
+        if coherence_rate < 60:
+            print(f"   - Low coherence rate: {coherence_rate:.1f}%")
+        if avg_perplexity >= 100:
+            print(f"   - High average perplexity: {avg_perplexity:.2f}")
+        if anomalies > 0:
+            print(f"   - Weight anomalies: {anomalies}")
+    
+    del model
+    torch.cuda.empty_cache()
+    return success
+
+def main():
+    print("="*60)
+    print("Stage 2: Simple Layer Duplication")
+    print("64 layers → 80 layers")
+    print("="*60)
+    
+    # Load weights from stage 1
+    print(f"\n📥 Loading model from: {INPUT_DIR}")
+    weights = load_model_sharted(INPUT_DIR)
+    
+    print(f"\n📊 Loaded {len(weights)} tensors")
+    
+    # Create new weight dictionary
+    new_weights = OrderedDict()
+    
+    # Copy non-layer weights
+    print("\n📋 Copying non-layer weights...")
+    for name, tensor in weights.items():
+        if not name.startswith("model.layers."):
+            new_weights[name] = tensor.clone()
+    
+    # Layer expansion with progress bar
+    print("\n🔄 Expanding layers with simple duplication...")
+    print("   Layers 0-23: Direct copy")
+    print("   Layers 24-39: Each layer duplicated once")
+    print("   Layers 40-63: Direct copy (shifted to 56-79)")
+    
+    new_layer_idx = 0
+    
+    with tqdm(total=TARGET_LAYERS, desc="Creating layers") as pbar:
+        # Copy layers 0-23 unchanged
+        for old_idx in range(24):
+            layer_weights = extract_layer_weights(weights, old_idx)
+            new_weights.update(create_layer_weights(layer_weights, new_layer_idx))
+            new_layer_idx += 1
+            pbar.update(1)
+        
+        # Duplicate layers 24-39
+        for old_idx in range(24, 40):
+            # Copy original layer
+            layer_weights = extract_layer_weights(weights, old_idx)
+            new_weights.update(create_layer_weights(layer_weights, new_layer_idx))
+            new_layer_idx += 1
+            pbar.update(1)
+            
+            # Duplicate the same layer
+            print(f"\n   Duplicating layer {old_idx} → layer {new_layer_idx}")
+            new_weights.update(create_layer_weights(layer_weights, new_layer_idx))
+            new_layer_idx += 1
+            pbar.update(1)
+        
+        # Copy layers 40-63 to positions 56-79
+        for old_idx in range(40, 64):
+            layer_weights = extract_layer_weights(weights, old_idx)
+            new_weights.update(create_layer_weights(layer_weights, new_layer_idx))
+            new_layer_idx += 1
+            pbar.update(1)
+    
+    print(f"\n✓ Created {new_layer_idx} layers")
+    
+    # Verify we have all layers
+    if new_layer_idx != TARGET_LAYERS:
+        print(f"\n❌ ERROR: Created {new_layer_idx} layers but expected {TARGET_LAYERS}")
+        print("Layer creation failed. Exiting.")
+        return False
+    
+    # Update config
+    print("\n📝 Updating model configuration...")
+    config_path = os.path.join(INPUT_DIR, "config.json")
+    with open(config_path, 'r') as f:
+        config = json.load(f)
+    
+    config['num_hidden_layers'] = TARGET_LAYERS
+    
+    # Save everything
+    print(f"\n💾 Saving expanded model to: {OUTPUT_DIR}")
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+    
+    # Save config
+    with open(os.path.join(OUTPUT_DIR, "config.json"), 'w') as f:
+        json.dump(config, f, indent=2)
+    
+    # Copy tokenizer files
+    tokenizer_files = [
+        'tokenizer.json', 'tokenizer_config.json', 
+        'special_tokens_map.json', 'generation_config.json'
+    ]
+    
+    for file in tokenizer_files:
+        src = os.path.join(INPUT_DIR, file)
+        dst = os.path.join(OUTPUT_DIR, file)
+        if os.path.exists(src):
+            shutil.copy(src, dst)
+    
+    # Save weights in sharted format
+    save_model_sharted(new_weights, OUTPUT_DIR)
+    
+    # Save metadata
+    metadata = {
+        "stage": "2-duplicate",
+        "source_model": INPUT_DIR,
+        "method": "Simple layer duplication",
+        "layer_mapping": {
+            "0-23": "0-23 (unchanged)",
+            "24-39": "24-55 (each duplicated once)",
+            "40-63": "56-79 (unchanged)"
+        },
+        "duplication_info": {
+            "method": "exact_copy",
+            "layers_duplicated": list(range(24, 40))
+        },
+        "final_layers": TARGET_LAYERS
+    }
+    
+    with open(os.path.join(OUTPUT_DIR, "stage2_metadata.json"), 'w') as f:
+        json.dump(metadata, f, indent=2)
+    
+    print("\n✅ Stage 2 duplication complete!")
+    
+    # Quick verification
+    print("\n🔍 Quick verification:")
+    print(f"   Total weights: {len(new_weights)}")
+    
+    # Count layers
+    layer_count = 0
+    for name in new_weights.keys():
+        if name.startswith("model.layers.") and ".input_layernorm.weight" in name:
+            layer_count += 1
+    
+    print(f"   Layer count: {layer_count} (expected: {TARGET_LAYERS})")
+    
+    # Check duplicate similarity
+    print("\n🔬 Checking layer duplication:")
+    test_component = "self_attn.q_proj.weight"
+    
+    # Check first duplicate pair
+    if f"model.layers.24.{test_component}" in new_weights and f"model.layers.25.{test_component}" in new_weights:
+        layer24 = new_weights[f"model.layers.24.{test_component}"]
+        layer25 = new_weights[f"model.layers.25.{test_component}"]
+        
+        # Should be identical
+        if torch.equal(layer24, layer25):
+            print("   ✓ Layer 24 and 25 are identical (as expected)")
+        else:
+            print("   ⚠️  Layer 24 and 25 differ (unexpected!)")
+    
+    print(f"\n🎉 SUCCESS! Model expanded to {TARGET_LAYERS} layers.")
+    print(f"📁 Output saved to: {OUTPUT_DIR}")
+    
+    # Run full diagnostics
+    arch_ok = verify_architecture(OUTPUT_DIR)
+    diag_ok = run_diagnostics(OUTPUT_DIR)
+    
+    if arch_ok and diag_ok:
+        print("\n🎊 FINAL SUCCESS! Your Qwen3-72B-DupeLayers model is ready and verified!")
+        print("\n📐 Final architecture:")
+        print("   Hidden size: 8192")
+        print("   Intermediate size: 29568") 
+        print("   Attention heads: 64")
+        print("   KV heads: 8")
+        print("   Layers: 80")
+        print("   Vocabulary: 151936")
+        print("\n💡 The model has passed all quality checks and is ready for use!")
+    else:
+        print("\n⚠️  Some verification issues detected. Please review the diagnostics above.")
+    
+    return arch_ok and diag_ok
+
+if __name__ == "__main__":
+    success = main()
+    exit(0 if success else 1)

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
+size 11422654

tokenizer_config.json

@@ -0,0 +1,240 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0].role == 'system' %}\n        {{- messages[0].content + '\\n\\n' }}\n    {%- endif %}\n    {{- \"# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0].role == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}\n{%- for message in messages[::-1] %}\n    {%- set index = (messages|length - 1) - loop.index0 %}\n    {%- if ns.multi_step_tool and message.role == \"user\" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}\n        {%- set ns.multi_step_tool = false %}\n        {%- set ns.last_query_index = index %}\n    {%- endif %}\n{%- endfor %}\n{%- for message in messages %}\n    {%- if message.content is string %}\n        {%- set content = message.content %}\n    {%- else %}\n        {%- set content = '' %}\n    {%- endif %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) %}\n        {{- '<|im_start|>' + message.role + '\\n' + content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {%- set reasoning_content = '' %}\n        {%- if message.reasoning_content is string %}\n            {%- set reasoning_content = message.reasoning_content %}\n        {%- else %}\n            {%- if '</think>' in content %}\n                {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n                {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n            {%- endif %}\n        {%- endif %}\n        {%- if loop.index0 > ns.last_query_index %}\n            {%- if loop.last or (not loop.last and reasoning_content) %}\n                {{- '<|im_start|>' + message.role + '\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') }}\n            {%- else %}\n                {{- '<|im_start|>' + message.role + '\\n' + content }}\n            {%- endif %}\n        {%- else %}\n            {{- '<|im_start|>' + message.role + '\\n' + content }}\n        {%- endif %}\n        {%- if message.tool_calls %}\n            {%- for tool_call in message.tool_calls %}\n                {%- if (loop.first and content) or (not loop.first) %}\n                    {{- '\\n' }}\n                {%- endif %}\n                {%- if tool_call.function %}\n                    {%- set tool_call = tool_call.function %}\n                {%- endif %}\n                {{- '<tool_call>\\n{\"name\": \"' }}\n                {{- tool_call.name }}\n                {{- '\", \"arguments\": ' }}\n                {%- if tool_call.arguments is string %}\n                    {{- tool_call.arguments }}\n                {%- else %}\n                    {{- tool_call.arguments | tojson }}\n                {%- endif %}\n                {{- '}\\n</tool_call>' }}\n            {%- endfor %}\n        {%- endif %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if loop.first or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n    {%- if enable_thinking is defined and enable_thinking is false %}\n        {{- '<think>\\n\\n</think>\\n\\n' }}\n    {%- endif %}\n{%- endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}