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 +1,26 @@
+{
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 131072,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 76,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.54.1",
+  "use_cache": false,
+  "vocab_size": 131072
+}

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "transformers_version": "4.54.1",
+  "use_cache": false
+}

latest

@@ -0,0 +1 @@
+global_step160

model-00001-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbbe6b36c1b4ab0ac40a675063ee9545b8d9c7023f109f9a92da5cf81e702aa6
+size 4865522496

model-00002-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03b82fde444cf1fa9ed57eac4e1b45a9d109d2f766cb7f6a36e50ffe1b3a19c8
+size 4907529424

model-00003-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21f3222f928af04de58211d9c17c7d9b80e06096fef01b898dc8a36bec702bb9
+size 4907529456

model-00004-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9416b62bbc030fd32681b4a7efd9706775ea05642ded343555baf5ec9f634be8
+size 4907529456

model-00005-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f586981176629bc7ab840fed793eb98d0cb97ee7f4d33cc8191e28d84059f71
+size 4907529456

model-00006-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc4c91d1174a32d90f3d1cad84ef782b3421584282af5b6fa452d9b6f25fb925
+size 4907529456

model-00007-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b81a309640bd66cab4980de8f702a0e2ca392023dbb5692693f82d362609d5ca
+size 4907529456

model-00008-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b03f65ae53839884fe6e98516ff9701e4ba22569f34b2468a7ebc01c7652157a
+size 4907529456

model-00009-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22a4e2f132d35f5a272dbe2f9d189e7552661d2233268f1ff067cb9c1d88022d
+size 4907496272

model.safetensors.index.json

@@ -0,0 +1,695 @@
+{
+  "metadata": {
+    "total_parameters": 783360,
+    "total_size": 44125644800
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00009-of-00009.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00009.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.input_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.mlp.down_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.mlp.up_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.post_attention_layernorm.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.32.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.33.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.33.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.33.mlp.gate_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.33.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.33.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.33.self_attn.k_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.33.self_attn.o_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.33.self_attn.q_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.33.self_attn.v_proj.weight": "model-00004-of-00009.safetensors",
+    "model.layers.34.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.34.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.35.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.36.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.37.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.38.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.39.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.40.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.40.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.input_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.mlp.down_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.mlp.up_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.post_attention_layernorm.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.41.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.42.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.42.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.42.mlp.gate_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.42.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.42.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.42.self_attn.k_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.42.self_attn.o_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.42.self_attn.q_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.42.self_attn.v_proj.weight": "model-00005-of-00009.safetensors",
+    "model.layers.43.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.43.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.44.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.45.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.46.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.47.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.48.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.49.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.50.input_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.mlp.down_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.mlp.up_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.post_attention_layernorm.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.50.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.51.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.51.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.51.mlp.gate_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.51.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.51.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.51.self_attn.k_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.51.self_attn.o_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.51.self_attn.q_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.51.self_attn.v_proj.weight": "model-00006-of-00009.safetensors",
+    "model.layers.52.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.52.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.53.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.54.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.55.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.56.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.57.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.58.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.input_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.mlp.down_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.mlp.up_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.post_attention_layernorm.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.59.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00009.safetensors",
+    "model.layers.60.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.60.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.60.mlp.gate_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.60.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.60.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.60.self_attn.k_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.60.self_attn.o_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.60.self_attn.q_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.60.self_attn.v_proj.weight": "model-00007-of-00009.safetensors",
+    "model.layers.61.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.61.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.62.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.63.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.64.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.65.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.66.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.67.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.input_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.mlp.down_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.mlp.up_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.post_attention_layernorm.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.68.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.69.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.69.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.69.mlp.gate_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.69.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.69.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.69.self_attn.k_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.69.self_attn.o_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.69.self_attn.q_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.69.self_attn.v_proj.weight": "model-00008-of-00009.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.70.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.70.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.71.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.72.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.73.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.74.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.input_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.mlp.down_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.mlp.gate_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.mlp.up_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.post_attention_layernorm.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.self_attn.k_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.self_attn.o_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.self_attn.q_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.75.self_attn.v_proj.weight": "model-00009-of-00009.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00009.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00009.safetensors",
+    "model.norm.weight": "model-00009-of-00009.safetensors"
+  }
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc7a5d9ca652a5bd1d04349d77fd0f56f8f1775e1b00b5cc5176468e017ac92a
+size 16389

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bd450366f4274a04e6d5dd0754dcf816f88f65b85ed9c88a1d081c89aada859
+size 16389

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a77d0583487d8cfff2e76df0532730c5485b42767b1ed8f4969f22094d840e3
+size 16389

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfc063696020bbba4af73e5ff59ec7700b0b1cdad96c500dc21ee7ad1944a49f
+size 16389

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6eef58bc47946544af1ba531a8c236e1942654ee216ea13ace82cce27989f69
+size 16389

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1817a933eaa47c02483f98f317c6922afa277826ec0331b0e7b317b1a450ede
+size 16389

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9879f2fc500918bc865d70f89366cd6d137a38668bd679637ef9fb550fc7ed7a
+size 16389

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6dbd587ebbf015d1ccd7042d123702b4ff325bb586cf67ebe78a6992b029aa9a
+size 16389

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c6a135c2fb72390422613b4a41730a83d024a52688dd46e22c1c928bd8711ed
+size 1465

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0240ce510f08e6c2041724e9043e33be9d251d1e4a4d94eb68cd47b954b61d2
+size 17078292

trainer_state.json

@@ -0,0 +1,1154 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9984399375975039,
+  "eval_steps": 500,
+  "global_step": 160,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0062402496099844,
+      "grad_norm": 10.74402992025825,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 2.0885,
+      "step": 1
+    },
+    {
+      "epoch": 0.0124804992199688,
+      "grad_norm": 10.414041822411015,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 2.0268,
+      "step": 2
+    },
+    {
+      "epoch": 0.0187207488299532,
+      "grad_norm": 9.463277664294495,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 2.0224,
+      "step": 3
+    },
+    {
+      "epoch": 0.0249609984399376,
+      "grad_norm": 3.816015682685396,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 2.0739,
+      "step": 4
+    },
+    {
+      "epoch": 0.031201248049921998,
+      "grad_norm": 2.45701507391519,
+      "learning_rate": 2e-06,
+      "loss": 2.0331,
+      "step": 5
+    },
+    {
+      "epoch": 0.0374414976599064,
+      "grad_norm": 2.2119224271122664,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 2.0191,
+      "step": 6
+    },
+    {
+      "epoch": 0.0436817472698908,
+      "grad_norm": 2.282301056375637,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 1.9805,
+      "step": 7
+    },
+    {
+      "epoch": 0.0499219968798752,
+      "grad_norm": 2.8111235551059672,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 2.0581,
+      "step": 8
+    },
+    {
+      "epoch": 0.056162246489859596,
+      "grad_norm": 2.290032450883113,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 2.0564,
+      "step": 9
+    },
+    {
+      "epoch": 0.062402496099843996,
+      "grad_norm": 2.7448772203230587,
+      "learning_rate": 4e-06,
+      "loss": 2.0477,
+      "step": 10
+    },
+    {
+      "epoch": 0.0686427457098284,
+      "grad_norm": 2.426609984220206,
+      "learning_rate": 4.4e-06,
+      "loss": 2.021,
+      "step": 11
+    },
+    {
+      "epoch": 0.0748829953198128,
+      "grad_norm": 2.8668649784307085,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 2.0695,
+      "step": 12
+    },
+    {
+      "epoch": 0.0811232449297972,
+      "grad_norm": 3.3242641740190484,
+      "learning_rate": 5.200000000000001e-06,
+      "loss": 2.0124,
+      "step": 13
+    },
+    {
+      "epoch": 0.0873634945397816,
+      "grad_norm": 2.61279395903613,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 2.1133,
+      "step": 14
+    },
+    {
+      "epoch": 0.093603744149766,
+      "grad_norm": 5.14184681992228,
+      "learning_rate": 6e-06,
+      "loss": 2.043,
+      "step": 15
+    },
+    {
+      "epoch": 0.0998439937597504,
+      "grad_norm": 2.8607405485409276,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 2.0186,
+      "step": 16
+    },
+    {
+      "epoch": 0.1060842433697348,
+      "grad_norm": 3.907631913383047,
+      "learning_rate": 6.8e-06,
+      "loss": 2.0141,
+      "step": 17
+    },
+    {
+      "epoch": 0.11232449297971919,
+      "grad_norm": 2.6838896552707014,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 2.0818,
+      "step": 18
+    },
+    {
+      "epoch": 0.11856474258970359,
+      "grad_norm": 3.6972529763342337,
+      "learning_rate": 7.6e-06,
+      "loss": 2.0148,
+      "step": 19
+    },
+    {
+      "epoch": 0.12480499219968799,
+      "grad_norm": 2.887987766429437,
+      "learning_rate": 8e-06,
+      "loss": 1.9952,
+      "step": 20
+    },
+    {
+      "epoch": 0.1310452418096724,
+      "grad_norm": 3.507471022279554,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 2.0801,
+      "step": 21
+    },
+    {
+      "epoch": 0.1372854914196568,
+      "grad_norm": 1.8196990054764854,
+      "learning_rate": 8.8e-06,
+      "loss": 2.0431,
+      "step": 22
+    },
+    {
+      "epoch": 0.1435257410296412,
+      "grad_norm": 4.634858392717085,
+      "learning_rate": 9.2e-06,
+      "loss": 2.0264,
+      "step": 23
+    },
+    {
+      "epoch": 0.1497659906396256,
+      "grad_norm": 3.62013301472126,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 2.0417,
+      "step": 24
+    },
+    {
+      "epoch": 0.15600624024961,
+      "grad_norm": 3.8817500707011305,
+      "learning_rate": 1e-05,
+      "loss": 2.0338,
+      "step": 25
+    },
+    {
+      "epoch": 0.1622464898595944,
+      "grad_norm": 4.087718448832571,
+      "learning_rate": 1.0400000000000002e-05,
+      "loss": 1.9754,
+      "step": 26
+    },
+    {
+      "epoch": 0.1684867394695788,
+      "grad_norm": 3.0490578592864828,
+      "learning_rate": 1.08e-05,
+      "loss": 2.0386,
+      "step": 27
+    },
+    {
+      "epoch": 0.1747269890795632,
+      "grad_norm": 2.9939386802357517,
+      "learning_rate": 1.1200000000000001e-05,
+      "loss": 2.0044,
+      "step": 28
+    },
+    {
+      "epoch": 0.1809672386895476,
+      "grad_norm": 3.935708629623511,
+      "learning_rate": 1.16e-05,
+      "loss": 2.0386,
+      "step": 29
+    },
+    {
+      "epoch": 0.187207488299532,
+      "grad_norm": 3.3632255605488686,
+      "learning_rate": 1.2e-05,
+      "loss": 2.0736,
+      "step": 30
+    },
+    {
+      "epoch": 0.1934477379095164,
+      "grad_norm": 4.075768572911316,
+      "learning_rate": 1.24e-05,
+      "loss": 2.0516,
+      "step": 31
+    },
+    {
+      "epoch": 0.1996879875195008,
+      "grad_norm": 1.9953301732913695,
+      "learning_rate": 1.2800000000000001e-05,
+      "loss": 2.0221,
+      "step": 32
+    },
+    {
+      "epoch": 0.2059282371294852,
+      "grad_norm": 5.026680509494266,
+      "learning_rate": 1.3200000000000002e-05,
+      "loss": 1.9922,
+      "step": 33
+    },
+    {
+      "epoch": 0.2121684867394696,
+      "grad_norm": 3.335169761696431,
+      "learning_rate": 1.36e-05,
+      "loss": 2.0422,
+      "step": 34
+    },
+    {
+      "epoch": 0.21840873634945399,
+      "grad_norm": 4.139988627533733,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 2.0286,
+      "step": 35
+    },
+    {
+      "epoch": 0.22464898595943839,
+      "grad_norm": 3.7359622657764002,
+      "learning_rate": 1.4400000000000001e-05,
+      "loss": 2.0034,
+      "step": 36
+    },
+    {
+      "epoch": 0.23088923556942278,
+      "grad_norm": 3.2432139656379233,
+      "learning_rate": 1.4800000000000002e-05,
+      "loss": 2.0811,
+      "step": 37
+    },
+    {
+      "epoch": 0.23712948517940718,
+      "grad_norm": 2.6586955034097532,
+      "learning_rate": 1.52e-05,
+      "loss": 2.02,
+      "step": 38
+    },
+    {
+      "epoch": 0.24336973478939158,
+      "grad_norm": 4.013453056484069,
+      "learning_rate": 1.5600000000000003e-05,
+      "loss": 1.9795,
+      "step": 39
+    },
+    {
+      "epoch": 0.24960998439937598,
+      "grad_norm": 2.521776704363988,
+      "learning_rate": 1.6e-05,
+      "loss": 1.9903,
+      "step": 40
+    },
+    {
+      "epoch": 0.25585023400936036,
+      "grad_norm": 4.065575685396291,
+      "learning_rate": 1.6400000000000002e-05,
+      "loss": 2.0221,
+      "step": 41
+    },
+    {
+      "epoch": 0.2620904836193448,
+      "grad_norm": 3.5403076970966643,
+      "learning_rate": 1.6800000000000002e-05,
+      "loss": 2.0709,
+      "step": 42
+    },
+    {
+      "epoch": 0.26833073322932915,
+      "grad_norm": 3.702151042696706,
+      "learning_rate": 1.72e-05,
+      "loss": 2.0208,
+      "step": 43
+    },
+    {
+      "epoch": 0.2745709828393136,
+      "grad_norm": 2.591541370214587,
+      "learning_rate": 1.76e-05,
+      "loss": 2.0225,
+      "step": 44
+    },
+    {
+      "epoch": 0.28081123244929795,
+      "grad_norm": 3.486718721357557,
+      "learning_rate": 1.8e-05,
+      "loss": 2.0365,
+      "step": 45
+    },
+    {
+      "epoch": 0.2870514820592824,
+      "grad_norm": 4.953839109375147,
+      "learning_rate": 1.84e-05,
+      "loss": 2.0201,
+      "step": 46
+    },
+    {
+      "epoch": 0.29329173166926675,
+      "grad_norm": 3.293398181575717,
+      "learning_rate": 1.8800000000000003e-05,
+      "loss": 2.0134,
+      "step": 47
+    },
+    {
+      "epoch": 0.2995319812792512,
+      "grad_norm": 5.579089802583317,
+      "learning_rate": 1.9200000000000003e-05,
+      "loss": 1.9558,
+      "step": 48
+    },
+    {
+      "epoch": 0.30577223088923555,
+      "grad_norm": 4.544122172306774,
+      "learning_rate": 1.9600000000000002e-05,
+      "loss": 1.9454,
+      "step": 49
+    },
+    {
+      "epoch": 0.31201248049922,
+      "grad_norm": 5.4447638064914985,
+      "learning_rate": 2e-05,
+      "loss": 2.0016,
+      "step": 50
+    },
+    {
+      "epoch": 0.31825273010920435,
+      "grad_norm": 9.597372129578162,
+      "learning_rate": 1.9981668194317142e-05,
+      "loss": 1.9972,
+      "step": 51
+    },
+    {
+      "epoch": 0.3244929797191888,
+      "grad_norm": 27.057608263740534,
+      "learning_rate": 1.9963031423290204e-05,
+      "loss": 2.0419,
+      "step": 52
+    },
+    {
+      "epoch": 0.33073322932917315,
+      "grad_norm": 7.246050894562496,
+      "learning_rate": 1.994408201304753e-05,
+      "loss": 2.03,
+      "step": 53
+    },
+    {
+      "epoch": 0.3369734789391576,
+      "grad_norm": 11.339707951526051,
+      "learning_rate": 1.9924812030075188e-05,
+      "loss": 2.0708,
+      "step": 54
+    },
+    {
+      "epoch": 0.34321372854914195,
+      "grad_norm": 54.33866443144764,
+      "learning_rate": 1.9905213270142184e-05,
+      "loss": 2.1948,
+      "step": 55
+    },
+    {
+      "epoch": 0.3494539781591264,
+      "grad_norm": 10.82899390141888,
+      "learning_rate": 1.988527724665392e-05,
+      "loss": 2.1432,
+      "step": 56
+    },
+    {
+      "epoch": 0.35569422776911075,
+      "grad_norm": 4.067023362710712,
+      "learning_rate": 1.986499517839923e-05,
+      "loss": 2.0887,
+      "step": 57
+    },
+    {
+      "epoch": 0.3619344773790952,
+      "grad_norm": 5.822340342195218,
+      "learning_rate": 1.9844357976653696e-05,
+      "loss": 2.0353,
+      "step": 58
+    },
+    {
+      "epoch": 0.36817472698907955,
+      "grad_norm": 3.5192080522606117,
+      "learning_rate": 1.9823356231599608e-05,
+      "loss": 2.0402,
+      "step": 59
+    },
+    {
+      "epoch": 0.374414976599064,
+      "grad_norm": 4.0093539975984065,
+      "learning_rate": 1.9801980198019806e-05,
+      "loss": 2.0968,
+      "step": 60
+    },
+    {
+      "epoch": 0.38065522620904835,
+      "grad_norm": 3.248210484978779,
+      "learning_rate": 1.9780219780219784e-05,
+      "loss": 2.0698,
+      "step": 61
+    },
+    {
+      "epoch": 0.3868954758190328,
+      "grad_norm": 9.86816642625139,
+      "learning_rate": 1.9758064516129035e-05,
+      "loss": 2.1396,
+      "step": 62
+    },
+    {
+      "epoch": 0.39313572542901715,
+      "grad_norm": 6.803518728015235,
+      "learning_rate": 1.9735503560528995e-05,
+      "loss": 2.0915,
+      "step": 63
+    },
+    {
+      "epoch": 0.3993759750390016,
+      "grad_norm": 9.116753527737151,
+      "learning_rate": 1.9712525667351133e-05,
+      "loss": 2.1454,
+      "step": 64
+    },
+    {
+      "epoch": 0.40561622464898595,
+      "grad_norm": 5.696843981671945,
+      "learning_rate": 1.9689119170984456e-05,
+      "loss": 2.1245,
+      "step": 65
+    },
+    {
+      "epoch": 0.4118564742589704,
+      "grad_norm": 14.915888442988669,
+      "learning_rate": 1.96652719665272e-05,
+      "loss": 2.121,
+      "step": 66
+    },
+    {
+      "epoch": 0.41809672386895474,
+      "grad_norm": 96.78241254851471,
+      "learning_rate": 1.9640971488912355e-05,
+      "loss": 2.3768,
+      "step": 67
+    },
+    {
+      "epoch": 0.4243369734789392,
+      "grad_norm": 11.648041872276787,
+      "learning_rate": 1.961620469083156e-05,
+      "loss": 2.2285,
+      "step": 68
+    },
+    {
+      "epoch": 0.43057722308892354,
+      "grad_norm": 7.174165495387168,
+      "learning_rate": 1.9590958019375676e-05,
+      "loss": 2.1671,
+      "step": 69
+    },
+    {
+      "epoch": 0.43681747269890797,
+      "grad_norm": 6.48742758649308,
+      "learning_rate": 1.956521739130435e-05,
+      "loss": 2.1101,
+      "step": 70
+    },
+    {
+      "epoch": 0.44305772230889234,
+      "grad_norm": 3.8309613955567263,
+      "learning_rate": 1.953896816684962e-05,
+      "loss": 2.1285,
+      "step": 71
+    },
+    {
+      "epoch": 0.44929797191887677,
+      "grad_norm": 4.383193420557969,
+      "learning_rate": 1.9512195121951222e-05,
+      "loss": 2.1133,
+      "step": 72
+    },
+    {
+      "epoch": 0.45553822152886114,
+      "grad_norm": 3.9756337036474765,
+      "learning_rate": 1.9484882418812992e-05,
+      "loss": 2.05,
+      "step": 73
+    },
+    {
+      "epoch": 0.46177847113884557,
+      "grad_norm": 2.628853316257943,
+      "learning_rate": 1.9457013574660635e-05,
+      "loss": 2.0764,
+      "step": 74
+    },
+    {
+      "epoch": 0.46801872074882994,
+      "grad_norm": 3.9077882447399372,
+      "learning_rate": 1.942857142857143e-05,
+      "loss": 2.0569,
+      "step": 75
+    },
+    {
+      "epoch": 0.47425897035881437,
+      "grad_norm": 2.153564772831578,
+      "learning_rate": 1.9399538106235567e-05,
+      "loss": 1.9753,
+      "step": 76
+    },
+    {
+      "epoch": 0.48049921996879874,
+      "grad_norm": 3.2742987852933907,
+      "learning_rate": 1.9369894982497083e-05,
+      "loss": 2.0728,
+      "step": 77
+    },
+    {
+      "epoch": 0.48673946957878317,
+      "grad_norm": 3.0871656421066236,
+      "learning_rate": 1.9339622641509436e-05,
+      "loss": 2.0725,
+      "step": 78
+    },
+    {
+      "epoch": 0.49297971918876754,
+      "grad_norm": 4.974500703202643,
+      "learning_rate": 1.9308700834326583e-05,
+      "loss": 2.0574,
+      "step": 79
+    },
+    {
+      "epoch": 0.49921996879875197,
+      "grad_norm": 1.9848768489870379,
+      "learning_rate": 1.9277108433734944e-05,
+      "loss": 2.039,
+      "step": 80
+    },
+    {
+      "epoch": 0.5054602184087363,
+      "grad_norm": 3.675535765562546,
+      "learning_rate": 1.9244823386114496e-05,
+      "loss": 2.0498,
+      "step": 81
+    },
+    {
+      "epoch": 0.5117004680187207,
+      "grad_norm": 3.646020776409229,
+      "learning_rate": 1.9211822660098524e-05,
+      "loss": 2.09,
+      "step": 82
+    },
+    {
+      "epoch": 0.5179407176287052,
+      "grad_norm": 3.2885099554782,
+      "learning_rate": 1.9178082191780822e-05,
+      "loss": 2.0594,
+      "step": 83
+    },
+    {
+      "epoch": 0.5241809672386896,
+      "grad_norm": 3.86257607313755,
+      "learning_rate": 1.9143576826196474e-05,
+      "loss": 2.0982,
+      "step": 84
+    },
+    {
+      "epoch": 0.5304212168486739,
+      "grad_norm": 2.9398571293411337,
+      "learning_rate": 1.9108280254777068e-05,
+      "loss": 2.0192,
+      "step": 85
+    },
+    {
+      "epoch": 0.5366614664586583,
+      "grad_norm": 5.691015626741047,
+      "learning_rate": 1.907216494845361e-05,
+      "loss": 2.0719,
+      "step": 86
+    },
+    {
+      "epoch": 0.5429017160686428,
+      "grad_norm": 3.2393091526236204,
+      "learning_rate": 1.9035202086049548e-05,
+      "loss": 2.0201,
+      "step": 87
+    },
+    {
+      "epoch": 0.5491419656786272,
+      "grad_norm": 7.878611924949738,
+      "learning_rate": 1.899736147757256e-05,
+      "loss": 2.0513,
+      "step": 88
+    },
+    {
+      "epoch": 0.5553822152886115,
+      "grad_norm": 5.586452362544675,
+      "learning_rate": 1.895861148197597e-05,
+      "loss": 2.008,
+      "step": 89
+    },
+    {
+      "epoch": 0.5616224648985959,
+      "grad_norm": 6.366113686688125,
+      "learning_rate": 1.891891891891892e-05,
+      "loss": 2.0222,
+      "step": 90
+    },
+    {
+      "epoch": 0.5678627145085804,
+      "grad_norm": 5.5469754048863305,
+      "learning_rate": 1.887824897400821e-05,
+      "loss": 2.0377,
+      "step": 91
+    },
+    {
+      "epoch": 0.5741029641185648,
+      "grad_norm": 4.927264399708736,
+      "learning_rate": 1.883656509695291e-05,
+      "loss": 2.0055,
+      "step": 92
+    },
+    {
+      "epoch": 0.5803432137285491,
+      "grad_norm": 4.400828802773919,
+      "learning_rate": 1.8793828892005612e-05,
+      "loss": 2.0385,
+      "step": 93
+    },
+    {
+      "epoch": 0.5865834633385335,
+      "grad_norm": 5.1654734914491,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 2.033,
+      "step": 94
+    },
+    {
+      "epoch": 0.592823712948518,
+      "grad_norm": 2.8256214952031224,
+      "learning_rate": 1.8705035971223024e-05,
+      "loss": 2.0456,
+      "step": 95
+    },
+    {
+      "epoch": 0.5990639625585024,
+      "grad_norm": 5.945292262432217,
+      "learning_rate": 1.8658892128279884e-05,
+      "loss": 2.0766,
+      "step": 96
+    },
+    {
+      "epoch": 0.6053042121684867,
+      "grad_norm": 3.1462260207166923,
+      "learning_rate": 1.861152141802068e-05,
+      "loss": 2.0502,
+      "step": 97
+    },
+    {
+      "epoch": 0.6115444617784711,
+      "grad_norm": 6.002424537736364,
+      "learning_rate": 1.8562874251497008e-05,
+      "loss": 2.0255,
+      "step": 98
+    },
+    {
+      "epoch": 0.6177847113884556,
+      "grad_norm": 4.92987161539047,
+      "learning_rate": 1.851289833080425e-05,
+      "loss": 2.0076,
+      "step": 99
+    },
+    {
+      "epoch": 0.62402496099844,
+      "grad_norm": 4.302040549632859,
+      "learning_rate": 1.846153846153846e-05,
+      "loss": 2.0523,
+      "step": 100
+    },
+    {
+      "epoch": 0.6302652106084243,
+      "grad_norm": 2.838409708113077,
+      "learning_rate": 1.840873634945398e-05,
+      "loss": 1.9646,
+      "step": 101
+    },
+    {
+      "epoch": 0.6365054602184087,
+      "grad_norm": 3.295108362363732,
+      "learning_rate": 1.8354430379746836e-05,
+      "loss": 1.9575,
+      "step": 102
+    },
+    {
+      "epoch": 0.6427457098283932,
+      "grad_norm": 2.2894649638697557,
+      "learning_rate": 1.8298555377207063e-05,
+      "loss": 1.9964,
+      "step": 103
+    },
+    {
+      "epoch": 0.6489859594383776,
+      "grad_norm": 3.7807867883237343,
+      "learning_rate": 1.8241042345276876e-05,
+      "loss": 1.9929,
+      "step": 104
+    },
+    {
+      "epoch": 0.6552262090483619,
+      "grad_norm": 3.5213953795099817,
+      "learning_rate": 1.8181818181818182e-05,
+      "loss": 2.0305,
+      "step": 105
+    },
+    {
+      "epoch": 0.6614664586583463,
+      "grad_norm": 3.180581118333888,
+      "learning_rate": 1.8120805369127517e-05,
+      "loss": 2.0,
+      "step": 106
+    },
+    {
+      "epoch": 0.6677067082683308,
+      "grad_norm": 3.7407765148739154,
+      "learning_rate": 1.8057921635434413e-05,
+      "loss": 1.9947,
+      "step": 107
+    },
+    {
+      "epoch": 0.6739469578783152,
+      "grad_norm": 2.4800887886136302,
+      "learning_rate": 1.7993079584775085e-05,
+      "loss": 1.945,
+      "step": 108
+    },
+    {
+      "epoch": 0.6801872074882995,
+      "grad_norm": 2.280159270411404,
+      "learning_rate": 1.792618629173989e-05,
+      "loss": 2.0461,
+      "step": 109
+    },
+    {
+      "epoch": 0.6864274570982839,
+      "grad_norm": 1.8103075862546911,
+      "learning_rate": 1.785714285714286e-05,
+      "loss": 2.0384,
+      "step": 110
+    },
+    {
+      "epoch": 0.6926677067082684,
+      "grad_norm": 2.7530066817239116,
+      "learning_rate": 1.7785843920145195e-05,
+      "loss": 2.0422,
+      "step": 111
+    },
+    {
+      "epoch": 0.6989079563182528,
+      "grad_norm": 2.385559179108808,
+      "learning_rate": 1.771217712177122e-05,
+      "loss": 2.0201,
+      "step": 112
+    },
+    {
+      "epoch": 0.7051482059282371,
+      "grad_norm": 3.4132861718816656,
+      "learning_rate": 1.7636022514071294e-05,
+      "loss": 2.0059,
+      "step": 113
+    },
+    {
+      "epoch": 0.7113884555382215,
+      "grad_norm": 1.9687493474305817,
+      "learning_rate": 1.755725190839695e-05,
+      "loss": 1.9424,
+      "step": 114
+    },
+    {
+      "epoch": 0.717628705148206,
+      "grad_norm": 3.3664419654172812,
+      "learning_rate": 1.7475728155339808e-05,
+      "loss": 1.9917,
+      "step": 115
+    },
+    {
+      "epoch": 0.7238689547581904,
+      "grad_norm": 1.7453289663664593,
+      "learning_rate": 1.7391304347826085e-05,
+      "loss": 1.9943,
+      "step": 116
+    },
+    {
+      "epoch": 0.7301092043681747,
+      "grad_norm": 3.513395200138233,
+      "learning_rate": 1.7303822937625752e-05,
+      "loss": 2.0497,
+      "step": 117
+    },
+    {
+      "epoch": 0.7363494539781591,
+      "grad_norm": 2.680135793492889,
+      "learning_rate": 1.721311475409836e-05,
+      "loss": 2.0534,
+      "step": 118
+    },
+    {
+      "epoch": 0.7425897035881436,
+      "grad_norm": 2.921520804571559,
+      "learning_rate": 1.711899791231733e-05,
+      "loss": 1.9642,
+      "step": 119
+    },
+    {
+      "epoch": 0.748829953198128,
+      "grad_norm": 2.3783552559811594,
+      "learning_rate": 1.7021276595744686e-05,
+      "loss": 2.0363,
+      "step": 120
+    },
+    {
+      "epoch": 0.7550702028081123,
+      "grad_norm": 3.4009796330846567,
+      "learning_rate": 1.691973969631237e-05,
+      "loss": 2.0112,
+      "step": 121
+    },
+    {
+      "epoch": 0.7613104524180967,
+      "grad_norm": 3.173059834064885,
+      "learning_rate": 1.6814159292035402e-05,
+      "loss": 2.036,
+      "step": 122
+    },
+    {
+      "epoch": 0.7675507020280812,
+      "grad_norm": 1.9816975235472727,
+      "learning_rate": 1.670428893905192e-05,
+      "loss": 2.0391,
+      "step": 123
+    },
+    {
+      "epoch": 0.7737909516380655,
+      "grad_norm": 2.3017687165242755,
+      "learning_rate": 1.6589861751152075e-05,
+      "loss": 2.0232,
+      "step": 124
+    },
+    {
+      "epoch": 0.7800312012480499,
+      "grad_norm": 2.094432126031965,
+      "learning_rate": 1.647058823529412e-05,
+      "loss": 2.0496,
+      "step": 125
+    },
+    {
+      "epoch": 0.7862714508580343,
+      "grad_norm": 4.369204228615871,
+      "learning_rate": 1.6346153846153847e-05,
+      "loss": 1.995,
+      "step": 126
+    },
+    {
+      "epoch": 0.7925117004680188,
+      "grad_norm": 2.707513968938277,
+      "learning_rate": 1.6216216216216215e-05,
+      "loss": 2.0909,
+      "step": 127
+    },
+    {
+      "epoch": 0.7987519500780031,
+      "grad_norm": 3.3723262775787437,
+      "learning_rate": 1.608040201005025e-05,
+      "loss": 2.0436,
+      "step": 128
+    },
+    {
+      "epoch": 0.8049921996879875,
+      "grad_norm": 4.542572647672642,
+      "learning_rate": 1.5938303341902316e-05,
+      "loss": 2.0467,
+      "step": 129
+    },
+    {
+      "epoch": 0.8112324492979719,
+      "grad_norm": 2.752970144346649,
+      "learning_rate": 1.5789473684210526e-05,
+      "loss": 2.026,
+      "step": 130
+    },
+    {
+      "epoch": 0.8174726989079563,
+      "grad_norm": 5.8547013314456695,
+      "learning_rate": 1.5633423180592992e-05,
+      "loss": 2.0379,
+      "step": 131
+    },
+    {
+      "epoch": 0.8237129485179407,
+      "grad_norm": 3.6188017739630824,
+      "learning_rate": 1.5469613259668508e-05,
+      "loss": 1.9989,
+      "step": 132
+    },
+    {
+      "epoch": 0.8299531981279251,
+      "grad_norm": 7.450963398151789,
+      "learning_rate": 1.529745042492918e-05,
+      "loss": 1.9937,
+      "step": 133
+    },
+    {
+      "epoch": 0.8361934477379095,
+      "grad_norm": 6.338885342053262,
+      "learning_rate": 1.5116279069767441e-05,
+      "loss": 2.0268,
+      "step": 134
+    },
+    {
+      "epoch": 0.8424336973478939,
+      "grad_norm": 4.068350711677805,
+      "learning_rate": 1.4925373134328359e-05,
+      "loss": 2.0657,
+      "step": 135
+    },
+    {
+      "epoch": 0.8486739469578783,
+      "grad_norm": 5.595793283481958,
+      "learning_rate": 1.4723926380368096e-05,
+      "loss": 2.0018,
+      "step": 136
+    },
+    {
+      "epoch": 0.8549141965678627,
+      "grad_norm": 2.188112379086696,
+      "learning_rate": 1.4511041009463722e-05,
+      "loss": 1.9973,
+      "step": 137
+    },
+    {
+      "epoch": 0.8611544461778471,
+      "grad_norm": 2.712340447399982,
+      "learning_rate": 1.4285714285714287e-05,
+      "loss": 1.9837,
+      "step": 138
+    },
+    {
+      "epoch": 0.8673946957878315,
+      "grad_norm": 4.524782432904714,
+      "learning_rate": 1.4046822742474917e-05,
+      "loss": 1.9805,
+      "step": 139
+    },
+    {
+      "epoch": 0.8736349453978159,
+      "grad_norm": 2.5304287757872026,
+      "learning_rate": 1.3793103448275862e-05,
+      "loss": 2.0088,
+      "step": 140
+    },
+    {
+      "epoch": 0.8798751950078003,
+      "grad_norm": 5.924048202005877,
+      "learning_rate": 1.3523131672597865e-05,
+      "loss": 1.9997,
+      "step": 141
+    },
+    {
+      "epoch": 0.8861154446177847,
+      "grad_norm": 5.465435613727891,
+      "learning_rate": 1.323529411764706e-05,
+      "loss": 1.9556,
+      "step": 142
+    },
+    {
+      "epoch": 0.8923556942277691,
+      "grad_norm": 3.404337115153408,
+      "learning_rate": 1.2927756653992395e-05,
+      "loss": 2.0268,
+      "step": 143
+    },
+    {
+      "epoch": 0.8985959438377535,
+      "grad_norm": 4.2172697858979165,
+      "learning_rate": 1.2598425196850392e-05,
+      "loss": 1.9997,
+      "step": 144
+    },
+    {
+      "epoch": 0.9048361934477379,
+      "grad_norm": 2.4532813229494255,
+      "learning_rate": 1.2244897959183674e-05,
+      "loss": 2.0326,
+      "step": 145
+    },
+    {
+      "epoch": 0.9110764430577223,
+      "grad_norm": 1.9619593022855364,
+      "learning_rate": 1.1864406779661018e-05,
+      "loss": 1.9498,
+      "step": 146
+    },
+    {
+      "epoch": 0.9173166926677067,
+      "grad_norm": 5.121507704839843,
+      "learning_rate": 1.1453744493392071e-05,
+      "loss": 1.9838,
+      "step": 147
+    },
+    {
+      "epoch": 0.9235569422776911,
+      "grad_norm": 4.5855609090084855,
+      "learning_rate": 1.1009174311926605e-05,
+      "loss": 1.9952,
+      "step": 148
+    },
+    {
+      "epoch": 0.9297971918876755,
+      "grad_norm": 2.8048531358016215,
+      "learning_rate": 1.0526315789473684e-05,
+      "loss": 1.9376,
+      "step": 149
+    },
+    {
+      "epoch": 0.9360374414976599,
+      "grad_norm": 3.1695682113009767,
+      "learning_rate": 1e-05,
+      "loss": 1.9892,
+      "step": 150
+    },
+    {
+      "epoch": 0.9422776911076443,
+      "grad_norm": 2.562068085601565,
+      "learning_rate": 9.424083769633508e-06,
+      "loss": 1.989,
+      "step": 151
+    },
+    {
+      "epoch": 0.9485179407176287,
+      "grad_norm": 2.5588451099258958,
+      "learning_rate": 8.79120879120879e-06,
+      "loss": 2.0055,
+      "step": 152
+    },
+    {
+      "epoch": 0.9547581903276131,
+      "grad_norm": 2.346708882799627,
+      "learning_rate": 8.092485549132947e-06,
+      "loss": 1.9344,
+      "step": 153
+    },
+    {
+      "epoch": 0.9609984399375975,
+      "grad_norm": 1.9275298796427967,
+      "learning_rate": 7.3170731707317065e-06,
+      "loss": 1.9777,
+      "step": 154
+    },
+    {
+      "epoch": 0.9672386895475819,
+      "grad_norm": 2.3144404690009934,
+      "learning_rate": 6.451612903225806e-06,
+      "loss": 2.0236,
+      "step": 155
+    },
+    {
+      "epoch": 0.9734789391575663,
+      "grad_norm": 2.0378918646209385,
+      "learning_rate": 5.479452054794519e-06,
+      "loss": 1.9846,
+      "step": 156
+    },
+    {
+      "epoch": 0.9797191887675507,
+      "grad_norm": 1.653128950918772,
+      "learning_rate": 4.379562043795621e-06,
+      "loss": 1.933,
+      "step": 157
+    },
+    {
+      "epoch": 0.9859594383775351,
+      "grad_norm": 1.5724161203243192,
+      "learning_rate": 3.125e-06,
+      "loss": 2.0316,
+      "step": 158
+    },
+    {
+      "epoch": 0.9921996879875195,
+      "grad_norm": 0.8954010032529827,
+      "learning_rate": 1.680672268907563e-06,
+      "loss": 1.9258,
+      "step": 159
+    },
+    {
+      "epoch": 0.9984399375975039,
+      "grad_norm": 1.0745203334736726,
+      "learning_rate": 0,
+      "loss": 2.0094,
+      "step": 160
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 160,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 80,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.0297691278239334e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f7f78a810b3e9165e4946aeb86b9fc41d44291554988961ec223e4638389bad
+size 9041

zero_to_fp32.py

@@ -0,0 +1,674 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)