model parts

config.json

@@ -0,0 +1,37 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 128000,
+  "eos_token_id": [
+    128001,
+    128008,
+    128009
+  ],
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 8192,
+  "mlp_bias": false,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 42,
+  "num_key_value_heads": 8,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "factor": 8.0,
+    "high_freq_factor": 4.0,
+    "low_freq_factor": 1.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "llama3"
+  },
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "transformers_version": "4.44.2",
+  "use_cache": true,
+  "vocab_size": 128256
+}

model-0000-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d637fa334db1c089e60841050bd4f043e0a80832e9e884370e2edab66bbfe62d
+size 2831225256

model-0001-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9f3c846034bf3d2a52a21dbe266399fd4d20ce4d2c647b4cbb2e45c5936d2c5
+size 1795232152

model-0002-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:114290c6c9f67363939d39342ab32e51392b67388ac4b944f4f89470f13afdd2
+size 2097222056

model-0003-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d8442722c9304af1149e0bd8e0d2ecb4dd5023d19a8af7e6bd340017614fc75a
+size 1778471336

model-0004-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a09d3a254fc9a1de942cf761e6f11a5ab556f78632e796169b75bf9c8361254
+size 1795232192

model-0005-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec40ef15a1f3a345c62ec275a248461b06e65a36b3f9e55edb44dedfd66eea10
+size 2097222080

model-0006-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55054a03fa1ac61f87bdeca7936fe07485a88f0f319d9f0cf15c347c504e4108
+size 1778471336

model-0007-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:744f921a95c9c66a0362908146996b0433d8ce61cc2e8c62cc1de045f6c35658
+size 1795232192

model-0008-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58116604b266a6c5ef70d37a2c35c57824d71d8df7e126c8b8ae579cd6c8772f
+size 2097222080

model-0009-of-0010.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6ec07209cc0d978308237d9bf93b891ebe9d82bb7abe9f4549a2c649d79c85a
+size 2359373336

model.safetensors.index.json

@@ -0,0 +1,389 @@
+{
+  "weight_map": {
+    "model.freqs_cis": "model-0000-of-0010.safetensors",
+    "model.embed_tokens.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-0000-of-0010.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-0000-of-0010.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-0001-of-0010.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-0001-of-0010.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-0002-of-0010.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-0003-of-0010.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-0003-of-0010.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-0004-of-0010.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-0004-of-0010.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-0005-of-0010.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-0006-of-0010.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-0006-of-0010.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-0007-of-0010.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-0007-of-0010.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-0008-of-0010.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-0009-of-0010.safetensors",
+    "model.norm.weight": "model-0009-of-0010.safetensors",
+    "lm_head.weight": "model-0009-of-0010.safetensors"
+  },
+  "metadata": {
+    "total_size": 20424859648
+  }
+}