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": 131074
+}

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_step238

model-00001-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e112fcf94cd6eaf57d73fd83d9a379cab050a732d82d75553c3eb5631797929e
+size 4865542976

model-00002-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06679e9923761d8d6d20bda79853ff0401600a2789325350611da43af3104974
+size 4907529424

model-00003-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d4f0275d745ab21c133f8c91fbbffd8610ffe0c5fb8106b5c2b0e68346f42e10
+size 4907529456

model-00004-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d494c411a764b038e3d27bd47a0a59c52f486df48550064bdea4e1ace93a36f8
+size 4907529456

model-00005-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ec9fd78224d891f0f42bc3e317ef0d20c6423b4ccdbef94c7aee08ae83e8804
+size 4907529456

model-00006-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ab5300e1c40c6b2dd2151285a8f8a1c03d389b8b86935ea97982c110552fd1e
+size 4907529456

model-00007-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b27b11bf3c890a1c5268775cbd9335e4b1b7a52cc49686ff06bfc4cde15ffa4b
+size 4907529456

model-00008-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c9e9b564deaf95b9f96edd1e07a28668f83224971a67d9a5e494677027115dc
+size 4907529456

model-00009-of-00009.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19423b8c78ed64e10bb55800c9cec5493f31f7a6bc86530067868476dcccdcec
+size 4907516752

model.safetensors.index.json

@@ -0,0 +1,695 @@
+{
+  "metadata": {
+    "total_parameters": 783360,
+    "total_size": 44125685760
+  },
+  "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:c13c68395a4d4274f2a718b92b364b9a25e9bf08268a05172b940cf97e9fa0f1
+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 sha256:48130f8c042761b84abbfbf10ad07efa7c26108a14e7a2a0402daa06e447a47a
+size 17078668

trainer_state.json

@@ -0,0 +1,1700 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 238,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.004201680672268907,
+      "grad_norm": 44.55869305414403,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 2.2685,
+      "step": 1
+    },
+    {
+      "epoch": 0.008403361344537815,
+      "grad_norm": 29.70889926040569,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 2.097,
+      "step": 2
+    },
+    {
+      "epoch": 0.012605042016806723,
+      "grad_norm": 21.840183292858224,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 2.156,
+      "step": 3
+    },
+    {
+      "epoch": 0.01680672268907563,
+      "grad_norm": 7.217101156991662,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 2.0766,
+      "step": 4
+    },
+    {
+      "epoch": 0.02100840336134454,
+      "grad_norm": 4.264808103375917,
+      "learning_rate": 2e-06,
+      "loss": 2.0865,
+      "step": 5
+    },
+    {
+      "epoch": 0.025210084033613446,
+      "grad_norm": 2.897016054882952,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 2.0046,
+      "step": 6
+    },
+    {
+      "epoch": 0.029411764705882353,
+      "grad_norm": 3.153455325913505,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 2.0024,
+      "step": 7
+    },
+    {
+      "epoch": 0.03361344537815126,
+      "grad_norm": 2.5168604497924743,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 2.1019,
+      "step": 8
+    },
+    {
+      "epoch": 0.037815126050420166,
+      "grad_norm": 1.9513097469453289,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 1.9845,
+      "step": 9
+    },
+    {
+      "epoch": 0.04201680672268908,
+      "grad_norm": 2.1379940411261926,
+      "learning_rate": 4e-06,
+      "loss": 2.0229,
+      "step": 10
+    },
+    {
+      "epoch": 0.046218487394957986,
+      "grad_norm": 3.3680679018237263,
+      "learning_rate": 4.4e-06,
+      "loss": 2.0471,
+      "step": 11
+    },
+    {
+      "epoch": 0.05042016806722689,
+      "grad_norm": 2.3075807669911166,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 1.9829,
+      "step": 12
+    },
+    {
+      "epoch": 0.0546218487394958,
+      "grad_norm": 2.8298274592428307,
+      "learning_rate": 5.200000000000001e-06,
+      "loss": 2.0004,
+      "step": 13
+    },
+    {
+      "epoch": 0.058823529411764705,
+      "grad_norm": 1.97265679203393,
+      "learning_rate": 5.600000000000001e-06,
+      "loss": 2.0155,
+      "step": 14
+    },
+    {
+      "epoch": 0.06302521008403361,
+      "grad_norm": 3.777203383008514,
+      "learning_rate": 6e-06,
+      "loss": 2.0398,
+      "step": 15
+    },
+    {
+      "epoch": 0.06722689075630252,
+      "grad_norm": 2.417303490372162,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 2.0173,
+      "step": 16
+    },
+    {
+      "epoch": 0.07142857142857142,
+      "grad_norm": 2.474079485633636,
+      "learning_rate": 6.8e-06,
+      "loss": 2.0312,
+      "step": 17
+    },
+    {
+      "epoch": 0.07563025210084033,
+      "grad_norm": 2.4565033464461825,
+      "learning_rate": 7.2000000000000005e-06,
+      "loss": 2.0186,
+      "step": 18
+    },
+    {
+      "epoch": 0.07983193277310924,
+      "grad_norm": 2.807819919133112,
+      "learning_rate": 7.6e-06,
+      "loss": 2.063,
+      "step": 19
+    },
+    {
+      "epoch": 0.08403361344537816,
+      "grad_norm": 3.3396601773601455,
+      "learning_rate": 8e-06,
+      "loss": 2.0507,
+      "step": 20
+    },
+    {
+      "epoch": 0.08823529411764706,
+      "grad_norm": 2.279073623922359,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 2.0199,
+      "step": 21
+    },
+    {
+      "epoch": 0.09243697478991597,
+      "grad_norm": 4.0027264668495395,
+      "learning_rate": 8.8e-06,
+      "loss": 2.0056,
+      "step": 22
+    },
+    {
+      "epoch": 0.09663865546218488,
+      "grad_norm": 2.1119183202477823,
+      "learning_rate": 9.2e-06,
+      "loss": 2.0131,
+      "step": 23
+    },
+    {
+      "epoch": 0.10084033613445378,
+      "grad_norm": 2.483816757482517,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 2.0412,
+      "step": 24
+    },
+    {
+      "epoch": 0.10504201680672269,
+      "grad_norm": 4.453875362500042,
+      "learning_rate": 1e-05,
+      "loss": 1.9927,
+      "step": 25
+    },
+    {
+      "epoch": 0.1092436974789916,
+      "grad_norm": 2.7338405519480102,
+      "learning_rate": 1.0400000000000002e-05,
+      "loss": 2.077,
+      "step": 26
+    },
+    {
+      "epoch": 0.1134453781512605,
+      "grad_norm": 3.2578386714830745,
+      "learning_rate": 1.08e-05,
+      "loss": 1.9791,
+      "step": 27
+    },
+    {
+      "epoch": 0.11764705882352941,
+      "grad_norm": 3.61499902295432,
+      "learning_rate": 1.1200000000000001e-05,
+      "loss": 2.0932,
+      "step": 28
+    },
+    {
+      "epoch": 0.12184873949579832,
+      "grad_norm": 2.405697489473297,
+      "learning_rate": 1.16e-05,
+      "loss": 2.0069,
+      "step": 29
+    },
+    {
+      "epoch": 0.12605042016806722,
+      "grad_norm": 3.7171276704442984,
+      "learning_rate": 1.2e-05,
+      "loss": 2.0175,
+      "step": 30
+    },
+    {
+      "epoch": 0.13025210084033614,
+      "grad_norm": 2.382345404566118,
+      "learning_rate": 1.24e-05,
+      "loss": 2.0293,
+      "step": 31
+    },
+    {
+      "epoch": 0.13445378151260504,
+      "grad_norm": 2.090570883694593,
+      "learning_rate": 1.2800000000000001e-05,
+      "loss": 1.9667,
+      "step": 32
+    },
+    {
+      "epoch": 0.13865546218487396,
+      "grad_norm": 2.3871999776031787,
+      "learning_rate": 1.3200000000000002e-05,
+      "loss": 1.984,
+      "step": 33
+    },
+    {
+      "epoch": 0.14285714285714285,
+      "grad_norm": 2.9172616267048297,
+      "learning_rate": 1.36e-05,
+      "loss": 1.9698,
+      "step": 34
+    },
+    {
+      "epoch": 0.14705882352941177,
+      "grad_norm": 3.6003653714969266,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 1.9858,
+      "step": 35
+    },
+    {
+      "epoch": 0.15126050420168066,
+      "grad_norm": 2.883664905666947,
+      "learning_rate": 1.4400000000000001e-05,
+      "loss": 2.0102,
+      "step": 36
+    },
+    {
+      "epoch": 0.15546218487394958,
+      "grad_norm": 3.1936897705363156,
+      "learning_rate": 1.4800000000000002e-05,
+      "loss": 1.9584,
+      "step": 37
+    },
+    {
+      "epoch": 0.15966386554621848,
+      "grad_norm": 4.8671768572506595,
+      "learning_rate": 1.52e-05,
+      "loss": 2.0283,
+      "step": 38
+    },
+    {
+      "epoch": 0.1638655462184874,
+      "grad_norm": 2.064543286389858,
+      "learning_rate": 1.5600000000000003e-05,
+      "loss": 2.0413,
+      "step": 39
+    },
+    {
+      "epoch": 0.16806722689075632,
+      "grad_norm": 5.795229466439302,
+      "learning_rate": 1.6e-05,
+      "loss": 1.953,
+      "step": 40
+    },
+    {
+      "epoch": 0.1722689075630252,
+      "grad_norm": 2.9275293275237093,
+      "learning_rate": 1.6400000000000002e-05,
+      "loss": 2.0297,
+      "step": 41
+    },
+    {
+      "epoch": 0.17647058823529413,
+      "grad_norm": 6.427773991759959,
+      "learning_rate": 1.6800000000000002e-05,
+      "loss": 2.037,
+      "step": 42
+    },
+    {
+      "epoch": 0.18067226890756302,
+      "grad_norm": 3.9938663346209666,
+      "learning_rate": 1.72e-05,
+      "loss": 2.0144,
+      "step": 43
+    },
+    {
+      "epoch": 0.18487394957983194,
+      "grad_norm": 6.86029152260702,
+      "learning_rate": 1.76e-05,
+      "loss": 2.04,
+      "step": 44
+    },
+    {
+      "epoch": 0.18907563025210083,
+      "grad_norm": 5.5731962014965255,
+      "learning_rate": 1.8e-05,
+      "loss": 2.0364,
+      "step": 45
+    },
+    {
+      "epoch": 0.19327731092436976,
+      "grad_norm": 12.208575068754369,
+      "learning_rate": 1.84e-05,
+      "loss": 1.996,
+      "step": 46
+    },
+    {
+      "epoch": 0.19747899159663865,
+      "grad_norm": 8.139261813550169,
+      "learning_rate": 1.8800000000000003e-05,
+      "loss": 1.9826,
+      "step": 47
+    },
+    {
+      "epoch": 0.20168067226890757,
+      "grad_norm": 3.964302435840079,
+      "learning_rate": 1.9200000000000003e-05,
+      "loss": 2.031,
+      "step": 48
+    },
+    {
+      "epoch": 0.20588235294117646,
+      "grad_norm": 8.405794668565285,
+      "learning_rate": 1.9600000000000002e-05,
+      "loss": 2.0434,
+      "step": 49
+    },
+    {
+      "epoch": 0.21008403361344538,
+      "grad_norm": 6.727672879110999,
+      "learning_rate": 2e-05,
+      "loss": 2.0017,
+      "step": 50
+    },
+    {
+      "epoch": 0.21428571428571427,
+      "grad_norm": 4.831959821672345,
+      "learning_rate": 1.998931052912881e-05,
+      "loss": 2.0233,
+      "step": 51
+    },
+    {
+      "epoch": 0.2184873949579832,
+      "grad_norm": 3.8023014702566456,
+      "learning_rate": 1.9978517722878624e-05,
+      "loss": 2.0239,
+      "step": 52
+    },
+    {
+      "epoch": 0.22268907563025211,
+      "grad_norm": 5.684333131264988,
+      "learning_rate": 1.9967620075553157e-05,
+      "loss": 2.0257,
+      "step": 53
+    },
+    {
+      "epoch": 0.226890756302521,
+      "grad_norm": 3.109833001328645,
+      "learning_rate": 1.995661605206074e-05,
+      "loss": 1.998,
+      "step": 54
+    },
+    {
+      "epoch": 0.23109243697478993,
+      "grad_norm": 4.006472777163142,
+      "learning_rate": 1.9945504087193465e-05,
+      "loss": 2.0066,
+      "step": 55
+    },
+    {
+      "epoch": 0.23529411764705882,
+      "grad_norm": 2.8517599823289275,
+      "learning_rate": 1.9934282584884995e-05,
+      "loss": 1.9921,
+      "step": 56
+    },
+    {
+      "epoch": 0.23949579831932774,
+      "grad_norm": 4.141928697362635,
+      "learning_rate": 1.992294991744634e-05,
+      "loss": 2.0385,
+      "step": 57
+    },
+    {
+      "epoch": 0.24369747899159663,
+      "grad_norm": 3.6225656174721,
+      "learning_rate": 1.991150442477876e-05,
+      "loss": 2.0633,
+      "step": 58
+    },
+    {
+      "epoch": 0.24789915966386555,
+      "grad_norm": 3.83078067702816,
+      "learning_rate": 1.9899944413563094e-05,
+      "loss": 2.0037,
+      "step": 59
+    },
+    {
+      "epoch": 0.25210084033613445,
+      "grad_norm": 3.1994001813904553,
+      "learning_rate": 1.9888268156424583e-05,
+      "loss": 2.0458,
+      "step": 60
+    },
+    {
+      "epoch": 0.25630252100840334,
+      "grad_norm": 3.6773836297404143,
+      "learning_rate": 1.9876473891072434e-05,
+      "loss": 1.976,
+      "step": 61
+    },
+    {
+      "epoch": 0.2605042016806723,
+      "grad_norm": 3.431265071184166,
+      "learning_rate": 1.9864559819413098e-05,
+      "loss": 1.9766,
+      "step": 62
+    },
+    {
+      "epoch": 0.2647058823529412,
+      "grad_norm": 3.9732233021275363,
+      "learning_rate": 1.9852524106636416e-05,
+      "loss": 2.0693,
+      "step": 63
+    },
+    {
+      "epoch": 0.2689075630252101,
+      "grad_norm": 2.424554270078444,
+      "learning_rate": 1.9840364880273663e-05,
+      "loss": 1.9929,
+      "step": 64
+    },
+    {
+      "epoch": 0.27310924369747897,
+      "grad_norm": 3.9615882262066004,
+      "learning_rate": 1.9828080229226363e-05,
+      "loss": 2.0031,
+      "step": 65
+    },
+    {
+      "epoch": 0.2773109243697479,
+      "grad_norm": 3.9063593348856185,
+      "learning_rate": 1.981566820276498e-05,
+      "loss": 2.0497,
+      "step": 66
+    },
+    {
+      "epoch": 0.2815126050420168,
+      "grad_norm": 2.5115336703536686,
+      "learning_rate": 1.980312680949624e-05,
+      "loss": 1.9918,
+      "step": 67
+    },
+    {
+      "epoch": 0.2857142857142857,
+      "grad_norm": 6.904884730461021,
+      "learning_rate": 1.9790454016298022e-05,
+      "loss": 2.0194,
+      "step": 68
+    },
+    {
+      "epoch": 0.28991596638655465,
+      "grad_norm": 5.150379742790806,
+      "learning_rate": 1.9777647747220598e-05,
+      "loss": 2.0278,
+      "step": 69
+    },
+    {
+      "epoch": 0.29411764705882354,
+      "grad_norm": 6.006210752025797,
+      "learning_rate": 1.9764705882352945e-05,
+      "loss": 2.0123,
+      "step": 70
+    },
+    {
+      "epoch": 0.29831932773109243,
+      "grad_norm": 5.198243582689913,
+      "learning_rate": 1.975162625665287e-05,
+      "loss": 1.9898,
+      "step": 71
+    },
+    {
+      "epoch": 0.3025210084033613,
+      "grad_norm": 4.853338232380691,
+      "learning_rate": 1.9738406658739597e-05,
+      "loss": 1.9936,
+      "step": 72
+    },
+    {
+      "epoch": 0.3067226890756303,
+      "grad_norm": 3.4564206889728353,
+      "learning_rate": 1.972504482964734e-05,
+      "loss": 1.9566,
+      "step": 73
+    },
+    {
+      "epoch": 0.31092436974789917,
+      "grad_norm": 6.3173822277231775,
+      "learning_rate": 1.971153846153846e-05,
+      "loss": 2.0204,
+      "step": 74
+    },
+    {
+      "epoch": 0.31512605042016806,
+      "grad_norm": 3.624893044720424,
+      "learning_rate": 1.9697885196374623e-05,
+      "loss": 2.0297,
+      "step": 75
+    },
+    {
+      "epoch": 0.31932773109243695,
+      "grad_norm": 7.519457183721233,
+      "learning_rate": 1.968408262454435e-05,
+      "loss": 2.008,
+      "step": 76
+    },
+    {
+      "epoch": 0.3235294117647059,
+      "grad_norm": 7.230947709467654,
+      "learning_rate": 1.967012828344533e-05,
+      "loss": 2.0154,
+      "step": 77
+    },
+    {
+      "epoch": 0.3277310924369748,
+      "grad_norm": 3.4232135075638537,
+      "learning_rate": 1.9656019656019657e-05,
+      "loss": 2.0078,
+      "step": 78
+    },
+    {
+      "epoch": 0.3319327731092437,
+      "grad_norm": 6.1448226278039435,
+      "learning_rate": 1.9641754169240276e-05,
+      "loss": 2.0825,
+      "step": 79
+    },
+    {
+      "epoch": 0.33613445378151263,
+      "grad_norm": 4.481660989035035,
+      "learning_rate": 1.9627329192546585e-05,
+      "loss": 2.0341,
+      "step": 80
+    },
+    {
+      "epoch": 0.3403361344537815,
+      "grad_norm": 3.4470653973720045,
+      "learning_rate": 1.961274203622736e-05,
+      "loss": 2.021,
+      "step": 81
+    },
+    {
+      "epoch": 0.3445378151260504,
+      "grad_norm": 4.100502642418945,
+      "learning_rate": 1.9597989949748744e-05,
+      "loss": 1.9552,
+      "step": 82
+    },
+    {
+      "epoch": 0.3487394957983193,
+      "grad_norm": 2.6080921726790143,
+      "learning_rate": 1.958307012002527e-05,
+      "loss": 1.9622,
+      "step": 83
+    },
+    {
+      "epoch": 0.35294117647058826,
+      "grad_norm": 4.002391242905452,
+      "learning_rate": 1.9567979669631518e-05,
+      "loss": 1.9698,
+      "step": 84
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 14.802641756278256,
+      "learning_rate": 1.955271565495208e-05,
+      "loss": 2.0274,
+      "step": 85
+    },
+    {
+      "epoch": 0.36134453781512604,
+      "grad_norm": 6.658990587554962,
+      "learning_rate": 1.9537275064267353e-05,
+      "loss": 1.9697,
+      "step": 86
+    },
+    {
+      "epoch": 0.36554621848739494,
+      "grad_norm": 4.172577334212283,
+      "learning_rate": 1.9521654815772465e-05,
+      "loss": 2.0092,
+      "step": 87
+    },
+    {
+      "epoch": 0.3697478991596639,
+      "grad_norm": 3.0733580825477125,
+      "learning_rate": 1.9505851755526658e-05,
+      "loss": 2.0093,
+      "step": 88
+    },
+    {
+      "epoch": 0.3739495798319328,
+      "grad_norm": 5.851398474672958,
+      "learning_rate": 1.9489862655330283e-05,
+      "loss": 1.9686,
+      "step": 89
+    },
+    {
+      "epoch": 0.37815126050420167,
+      "grad_norm": 2.766976890662351,
+      "learning_rate": 1.9473684210526315e-05,
+      "loss": 1.9972,
+      "step": 90
+    },
+    {
+      "epoch": 0.38235294117647056,
+      "grad_norm": 6.125408033621872,
+      "learning_rate": 1.945731303772336e-05,
+      "loss": 1.9977,
+      "step": 91
+    },
+    {
+      "epoch": 0.3865546218487395,
+      "grad_norm": 5.069952160426001,
+      "learning_rate": 1.9440745672436754e-05,
+      "loss": 2.025,
+      "step": 92
+    },
+    {
+      "epoch": 0.3907563025210084,
+      "grad_norm": 5.52702104127559,
+      "learning_rate": 1.9423978566644342e-05,
+      "loss": 2.013,
+      "step": 93
+    },
+    {
+      "epoch": 0.3949579831932773,
+      "grad_norm": 3.425995814770311,
+      "learning_rate": 1.9407008086253373e-05,
+      "loss": 2.0085,
+      "step": 94
+    },
+    {
+      "epoch": 0.39915966386554624,
+      "grad_norm": 5.726432596874525,
+      "learning_rate": 1.938983050847458e-05,
+      "loss": 2.0112,
+      "step": 95
+    },
+    {
+      "epoch": 0.40336134453781514,
+      "grad_norm": 5.6578297872521555,
+      "learning_rate": 1.9372442019099593e-05,
+      "loss": 1.9798,
+      "step": 96
+    },
+    {
+      "epoch": 0.40756302521008403,
+      "grad_norm": 3.9521170490922457,
+      "learning_rate": 1.935483870967742e-05,
+      "loss": 1.9871,
+      "step": 97
+    },
+    {
+      "epoch": 0.4117647058823529,
+      "grad_norm": 4.36120065424456,
+      "learning_rate": 1.9337016574585635e-05,
+      "loss": 2.001,
+      "step": 98
+    },
+    {
+      "epoch": 0.41596638655462187,
+      "grad_norm": 2.250844866874491,
+      "learning_rate": 1.9318971507991666e-05,
+      "loss": 2.0397,
+      "step": 99
+    },
+    {
+      "epoch": 0.42016806722689076,
+      "grad_norm": 6.54319081589665,
+      "learning_rate": 1.9300699300699302e-05,
+      "loss": 2.0261,
+      "step": 100
+    },
+    {
+      "epoch": 0.42436974789915966,
+      "grad_norm": 3.828229581804093,
+      "learning_rate": 1.9282195636875443e-05,
+      "loss": 2.0009,
+      "step": 101
+    },
+    {
+      "epoch": 0.42857142857142855,
+      "grad_norm": 7.878633920895527,
+      "learning_rate": 1.926345609065156e-05,
+      "loss": 2.0141,
+      "step": 102
+    },
+    {
+      "epoch": 0.4327731092436975,
+      "grad_norm": 6.260079646545892,
+      "learning_rate": 1.9244476122594442e-05,
+      "loss": 2.006,
+      "step": 103
+    },
+    {
+      "epoch": 0.4369747899159664,
+      "grad_norm": 7.322267536389183,
+      "learning_rate": 1.9225251076040173e-05,
+      "loss": 2.0253,
+      "step": 104
+    },
+    {
+      "epoch": 0.4411764705882353,
+      "grad_norm": 7.045810380338178,
+      "learning_rate": 1.92057761732852e-05,
+      "loss": 2.0272,
+      "step": 105
+    },
+    {
+      "epoch": 0.44537815126050423,
+      "grad_norm": 4.411630608320293,
+      "learning_rate": 1.918604651162791e-05,
+      "loss": 2.0419,
+      "step": 106
+    },
+    {
+      "epoch": 0.4495798319327731,
+      "grad_norm": 5.529174209678458,
+      "learning_rate": 1.9166057059253843e-05,
+      "loss": 1.9525,
+      "step": 107
+    },
+    {
+      "epoch": 0.453781512605042,
+      "grad_norm": 4.103353132247119,
+      "learning_rate": 1.9145802650957293e-05,
+      "loss": 1.9716,
+      "step": 108
+    },
+    {
+      "epoch": 0.4579831932773109,
+      "grad_norm": 3.1630552885322483,
+      "learning_rate": 1.9125277983691625e-05,
+      "loss": 2.031,
+      "step": 109
+    },
+    {
+      "epoch": 0.46218487394957986,
+      "grad_norm": 5.665264096518991,
+      "learning_rate": 1.9104477611940298e-05,
+      "loss": 1.9557,
+      "step": 110
+    },
+    {
+      "epoch": 0.46638655462184875,
+      "grad_norm": 3.775929121105562,
+      "learning_rate": 1.9083395942900076e-05,
+      "loss": 1.9564,
+      "step": 111
+    },
+    {
+      "epoch": 0.47058823529411764,
+      "grad_norm": 7.0736083628928474,
+      "learning_rate": 1.9062027231467476e-05,
+      "loss": 1.9665,
+      "step": 112
+    },
+    {
+      "epoch": 0.47478991596638653,
+      "grad_norm": 6.993527716006551,
+      "learning_rate": 1.9040365575019042e-05,
+      "loss": 1.9915,
+      "step": 113
+    },
+    {
+      "epoch": 0.4789915966386555,
+      "grad_norm": 4.281127299421095,
+      "learning_rate": 1.9018404907975462e-05,
+      "loss": 1.9976,
+      "step": 114
+    },
+    {
+      "epoch": 0.4831932773109244,
+      "grad_norm": 3.995816394994801,
+      "learning_rate": 1.8996138996139e-05,
+      "loss": 2.0068,
+      "step": 115
+    },
+    {
+      "epoch": 0.48739495798319327,
+      "grad_norm": 4.989712338733302,
+      "learning_rate": 1.8973561430793158e-05,
+      "loss": 1.998,
+      "step": 116
+    },
+    {
+      "epoch": 0.49159663865546216,
+      "grad_norm": 2.9839440466434075,
+      "learning_rate": 1.895066562255286e-05,
+      "loss": 1.9432,
+      "step": 117
+    },
+    {
+      "epoch": 0.4957983193277311,
+      "grad_norm": 6.46520187654929,
+      "learning_rate": 1.8927444794952684e-05,
+      "loss": 2.0337,
+      "step": 118
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 4.640602987227591,
+      "learning_rate": 1.890389197776013e-05,
+      "loss": 1.9999,
+      "step": 119
+    },
+    {
+      "epoch": 0.5042016806722689,
+      "grad_norm": 6.059899284979291,
+      "learning_rate": 1.8880000000000002e-05,
+      "loss": 2.0213,
+      "step": 120
+    },
+    {
+      "epoch": 0.5084033613445378,
+      "grad_norm": 5.569560355211523,
+      "learning_rate": 1.8855761482675264e-05,
+      "loss": 1.9409,
+      "step": 121
+    },
+    {
+      "epoch": 0.5126050420168067,
+      "grad_norm": 6.311272641039366,
+      "learning_rate": 1.8831168831168833e-05,
+      "loss": 2.028,
+      "step": 122
+    },
+    {
+      "epoch": 0.5168067226890757,
+      "grad_norm": 6.15178203152809,
+      "learning_rate": 1.8806214227309895e-05,
+      "loss": 1.9764,
+      "step": 123
+    },
+    {
+      "epoch": 0.5210084033613446,
+      "grad_norm": 2.0799803021864247,
+      "learning_rate": 1.8780889621087317e-05,
+      "loss": 2.0007,
+      "step": 124
+    },
+    {
+      "epoch": 0.5252100840336135,
+      "grad_norm": 6.678022469624418,
+      "learning_rate": 1.8755186721991702e-05,
+      "loss": 2.0298,
+      "step": 125
+    },
+    {
+      "epoch": 0.5294117647058824,
+      "grad_norm": 4.809492334867139,
+      "learning_rate": 1.8729096989966558e-05,
+      "loss": 1.9669,
+      "step": 126
+    },
+    {
+      "epoch": 0.5336134453781513,
+      "grad_norm": 4.2168705479113076,
+      "learning_rate": 1.870261162594777e-05,
+      "loss": 1.956,
+      "step": 127
+    },
+    {
+      "epoch": 0.5378151260504201,
+      "grad_norm": 4.435134241605622,
+      "learning_rate": 1.867572156196944e-05,
+      "loss": 1.9846,
+      "step": 128
+    },
+    {
+      "epoch": 0.542016806722689,
+      "grad_norm": 2.5224089090419755,
+      "learning_rate": 1.864841745081266e-05,
+      "loss": 1.9454,
+      "step": 129
+    },
+    {
+      "epoch": 0.5462184873949579,
+      "grad_norm": 3.437904301059768,
+      "learning_rate": 1.862068965517242e-05,
+      "loss": 2.022,
+      "step": 130
+    },
+    {
+      "epoch": 0.5504201680672269,
+      "grad_norm": 2.318004783983716,
+      "learning_rate": 1.8592528236316248e-05,
+      "loss": 1.9557,
+      "step": 131
+    },
+    {
+      "epoch": 0.5546218487394958,
+      "grad_norm": 3.53317396981907,
+      "learning_rate": 1.8563922942206657e-05,
+      "loss": 1.9826,
+      "step": 132
+    },
+    {
+      "epoch": 0.5588235294117647,
+      "grad_norm": 2.7518300571506122,
+      "learning_rate": 1.853486319505737e-05,
+      "loss": 2.0515,
+      "step": 133
+    },
+    {
+      "epoch": 0.5630252100840336,
+      "grad_norm": 3.4652674934431653,
+      "learning_rate": 1.8505338078291815e-05,
+      "loss": 2.0036,
+      "step": 134
+    },
+    {
+      "epoch": 0.5672268907563025,
+      "grad_norm": 2.073139989482465,
+      "learning_rate": 1.8475336322869956e-05,
+      "loss": 1.9287,
+      "step": 135
+    },
+    {
+      "epoch": 0.5714285714285714,
+      "grad_norm": 2.3479784995361563,
+      "learning_rate": 1.8444846292947557e-05,
+      "loss": 2.0144,
+      "step": 136
+    },
+    {
+      "epoch": 0.5756302521008403,
+      "grad_norm": 2.2937651705984092,
+      "learning_rate": 1.8413855970829533e-05,
+      "loss": 1.9501,
+      "step": 137
+    },
+    {
+      "epoch": 0.5798319327731093,
+      "grad_norm": 2.253237599566215,
+      "learning_rate": 1.838235294117647e-05,
+      "loss": 1.9936,
+      "step": 138
+    },
+    {
+      "epoch": 0.5840336134453782,
+      "grad_norm": 4.617336783754218,
+      "learning_rate": 1.8350324374420762e-05,
+      "loss": 1.9965,
+      "step": 139
+    },
+    {
+      "epoch": 0.5882352941176471,
+      "grad_norm": 1.9433194866860843,
+      "learning_rate": 1.8317757009345797e-05,
+      "loss": 1.936,
+      "step": 140
+    },
+    {
+      "epoch": 0.592436974789916,
+      "grad_norm": 5.449953087750545,
+      "learning_rate": 1.8284637134778513e-05,
+      "loss": 2.0189,
+      "step": 141
+    },
+    {
+      "epoch": 0.5966386554621849,
+      "grad_norm": 3.568174133988601,
+      "learning_rate": 1.8250950570342207e-05,
+      "loss": 1.9742,
+      "step": 142
+    },
+    {
+      "epoch": 0.6008403361344538,
+      "grad_norm": 4.677011649909216,
+      "learning_rate": 1.821668264621285e-05,
+      "loss": 1.9878,
+      "step": 143
+    },
+    {
+      "epoch": 0.6050420168067226,
+      "grad_norm": 4.1362670214662,
+      "learning_rate": 1.8181818181818182e-05,
+      "loss": 1.9743,
+      "step": 144
+    },
+    {
+      "epoch": 0.6092436974789915,
+      "grad_norm": 3.8237728341672637,
+      "learning_rate": 1.8146341463414637e-05,
+      "loss": 1.9885,
+      "step": 145
+    },
+    {
+      "epoch": 0.6134453781512605,
+      "grad_norm": 2.8786113722601345,
+      "learning_rate": 1.811023622047244e-05,
+      "loss": 1.9818,
+      "step": 146
+    },
+    {
+      "epoch": 0.6176470588235294,
+      "grad_norm": 3.2546032317770965,
+      "learning_rate": 1.807348560079444e-05,
+      "loss": 1.9787,
+      "step": 147
+    },
+    {
+      "epoch": 0.6218487394957983,
+      "grad_norm": 2.4795374553303646,
+      "learning_rate": 1.8036072144288578e-05,
+      "loss": 1.9819,
+      "step": 148
+    },
+    {
+      "epoch": 0.6260504201680672,
+      "grad_norm": 4.353223204315405,
+      "learning_rate": 1.799797775530839e-05,
+      "loss": 2.0087,
+      "step": 149
+    },
+    {
+      "epoch": 0.6302521008403361,
+      "grad_norm": 2.536117436581857,
+      "learning_rate": 1.795918367346939e-05,
+      "loss": 1.9899,
+      "step": 150
+    },
+    {
+      "epoch": 0.634453781512605,
+      "grad_norm": 4.675342898713688,
+      "learning_rate": 1.791967044284243e-05,
+      "loss": 2.0093,
+      "step": 151
+    },
+    {
+      "epoch": 0.6386554621848739,
+      "grad_norm": 3.5893339455691375,
+      "learning_rate": 1.7879417879417883e-05,
+      "loss": 1.9744,
+      "step": 152
+    },
+    {
+      "epoch": 0.6428571428571429,
+      "grad_norm": 3.860674951683096,
+      "learning_rate": 1.7838405036726128e-05,
+      "loss": 1.985,
+      "step": 153
+    },
+    {
+      "epoch": 0.6470588235294118,
+      "grad_norm": 7.713909731021625,
+      "learning_rate": 1.779661016949153e-05,
+      "loss": 1.9906,
+      "step": 154
+    },
+    {
+      "epoch": 0.6512605042016807,
+      "grad_norm": 5.045657195392537,
+      "learning_rate": 1.7754010695187164e-05,
+      "loss": 1.9677,
+      "step": 155
+    },
+    {
+      "epoch": 0.6554621848739496,
+      "grad_norm": 4.841553822054665,
+      "learning_rate": 1.7710583153347732e-05,
+      "loss": 1.9795,
+      "step": 156
+    },
+    {
+      "epoch": 0.6596638655462185,
+      "grad_norm": 35.907518892108044,
+      "learning_rate": 1.766630316248637e-05,
+      "loss": 2.0343,
+      "step": 157
+    },
+    {
+      "epoch": 0.6638655462184874,
+      "grad_norm": 5.094598053404981,
+      "learning_rate": 1.7621145374449342e-05,
+      "loss": 1.9854,
+      "step": 158
+    },
+    {
+      "epoch": 0.6680672268907563,
+      "grad_norm": 5.241476192542538,
+      "learning_rate": 1.757508342602892e-05,
+      "loss": 2.0368,
+      "step": 159
+    },
+    {
+      "epoch": 0.6722689075630253,
+      "grad_norm": 3.5660591980220264,
+      "learning_rate": 1.7528089887640448e-05,
+      "loss": 2.0254,
+      "step": 160
+    },
+    {
+      "epoch": 0.6764705882352942,
+      "grad_norm": 22.06349699018364,
+      "learning_rate": 1.7480136208853575e-05,
+      "loss": 2.0075,
+      "step": 161
+    },
+    {
+      "epoch": 0.680672268907563,
+      "grad_norm": 109.95776153080195,
+      "learning_rate": 1.7431192660550456e-05,
+      "loss": 2.4162,
+      "step": 162
+    },
+    {
+      "epoch": 0.6848739495798319,
+      "grad_norm": 36.65714969339457,
+      "learning_rate": 1.738122827346466e-05,
+      "loss": 2.258,
+      "step": 163
+    },
+    {
+      "epoch": 0.6890756302521008,
+      "grad_norm": 10.595173271392715,
+      "learning_rate": 1.733021077283372e-05,
+      "loss": 2.0676,
+      "step": 164
+    },
+    {
+      "epoch": 0.6932773109243697,
+      "grad_norm": 4.562219758211684,
+      "learning_rate": 1.7278106508875744e-05,
+      "loss": 2.0168,
+      "step": 165
+    },
+    {
+      "epoch": 0.6974789915966386,
+      "grad_norm": 4.7820348463766855,
+      "learning_rate": 1.722488038277512e-05,
+      "loss": 2.0695,
+      "step": 166
+    },
+    {
+      "epoch": 0.7016806722689075,
+      "grad_norm": 6.498006110568237,
+      "learning_rate": 1.717049576783555e-05,
+      "loss": 2.0552,
+      "step": 167
+    },
+    {
+      "epoch": 0.7058823529411765,
+      "grad_norm": 4.147928665279015,
+      "learning_rate": 1.7114914425427874e-05,
+      "loss": 2.0532,
+      "step": 168
+    },
+    {
+      "epoch": 0.7100840336134454,
+      "grad_norm": 4.002117894930057,
+      "learning_rate": 1.7058096415327567e-05,
+      "loss": 2.0581,
+      "step": 169
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 3.5097444381611695,
+      "learning_rate": 1.7e-05,
+      "loss": 2.0217,
+      "step": 170
+    },
+    {
+      "epoch": 0.7184873949579832,
+      "grad_norm": 2.9351838452700356,
+      "learning_rate": 1.6940581542351452e-05,
+      "loss": 2.0257,
+      "step": 171
+    },
+    {
+      "epoch": 0.7226890756302521,
+      "grad_norm": 3.4681847547655082,
+      "learning_rate": 1.687979539641944e-05,
+      "loss": 1.982,
+      "step": 172
+    },
+    {
+      "epoch": 0.726890756302521,
+      "grad_norm": 3.4618786514761672,
+      "learning_rate": 1.681759379042691e-05,
+      "loss": 2.0102,
+      "step": 173
+    },
+    {
+      "epoch": 0.7310924369747899,
+      "grad_norm": 1.8014313595666087,
+      "learning_rate": 1.675392670157068e-05,
+      "loss": 1.9616,
+      "step": 174
+    },
+    {
+      "epoch": 0.7352941176470589,
+      "grad_norm": 3.7841060015955668,
+      "learning_rate": 1.6688741721854306e-05,
+      "loss": 1.9823,
+      "step": 175
+    },
+    {
+      "epoch": 0.7394957983193278,
+      "grad_norm": 3.4243966678626907,
+      "learning_rate": 1.6621983914209116e-05,
+      "loss": 1.9708,
+      "step": 176
+    },
+    {
+      "epoch": 0.7436974789915967,
+      "grad_norm": 2.317331955664329,
+      "learning_rate": 1.655359565807327e-05,
+      "loss": 2.0252,
+      "step": 177
+    },
+    {
+      "epoch": 0.7478991596638656,
+      "grad_norm": 2.3772843387701954,
+      "learning_rate": 1.6483516483516486e-05,
+      "loss": 1.9816,
+      "step": 178
+    },
+    {
+      "epoch": 0.7521008403361344,
+      "grad_norm": 2.918683255617424,
+      "learning_rate": 1.6411682892906817e-05,
+      "loss": 1.994,
+      "step": 179
+    },
+    {
+      "epoch": 0.7563025210084033,
+      "grad_norm": 3.635923773345395,
+      "learning_rate": 1.633802816901409e-05,
+      "loss": 2.0706,
+      "step": 180
+    },
+    {
+      "epoch": 0.7605042016806722,
+      "grad_norm": 3.8699476267227886,
+      "learning_rate": 1.6262482168330956e-05,
+      "loss": 2.0665,
+      "step": 181
+    },
+    {
+      "epoch": 0.7647058823529411,
+      "grad_norm": 1.9700257045356901,
+      "learning_rate": 1.6184971098265897e-05,
+      "loss": 2.0561,
+      "step": 182
+    },
+    {
+      "epoch": 0.7689075630252101,
+      "grad_norm": 5.73128438126555,
+      "learning_rate": 1.610541727672035e-05,
+      "loss": 1.9914,
+      "step": 183
+    },
+    {
+      "epoch": 0.773109243697479,
+      "grad_norm": 2.4583848791829097,
+      "learning_rate": 1.6023738872403564e-05,
+      "loss": 2.0165,
+      "step": 184
+    },
+    {
+      "epoch": 0.7773109243697479,
+      "grad_norm": 6.505054560240348,
+      "learning_rate": 1.5939849624060154e-05,
+      "loss": 1.9605,
+      "step": 185
+    },
+    {
+      "epoch": 0.7815126050420168,
+      "grad_norm": 5.1975649709508875,
+      "learning_rate": 1.5853658536585366e-05,
+      "loss": 2.0139,
+      "step": 186
+    },
+    {
+      "epoch": 0.7857142857142857,
+      "grad_norm": 5.343558620909131,
+      "learning_rate": 1.5765069551777436e-05,
+      "loss": 1.9229,
+      "step": 187
+    },
+    {
+      "epoch": 0.7899159663865546,
+      "grad_norm": 4.150079503246313,
+      "learning_rate": 1.567398119122257e-05,
+      "loss": 1.9482,
+      "step": 188
+    },
+    {
+      "epoch": 0.7941176470588235,
+      "grad_norm": 4.778873512920807,
+      "learning_rate": 1.5580286168521466e-05,
+      "loss": 1.984,
+      "step": 189
+    },
+    {
+      "epoch": 0.7983193277310925,
+      "grad_norm": 2.9128904162117135,
+      "learning_rate": 1.5483870967741936e-05,
+      "loss": 1.9957,
+      "step": 190
+    },
+    {
+      "epoch": 0.8025210084033614,
+      "grad_norm": 5.74489850121851,
+      "learning_rate": 1.5384615384615384e-05,
+      "loss": 1.9826,
+      "step": 191
+    },
+    {
+      "epoch": 0.8067226890756303,
+      "grad_norm": 3.2864350540006715,
+      "learning_rate": 1.5282392026578074e-05,
+      "loss": 2.0637,
+      "step": 192
+    },
+    {
+      "epoch": 0.8109243697478992,
+      "grad_norm": 6.277182987109688,
+      "learning_rate": 1.5177065767284993e-05,
+      "loss": 1.9541,
+      "step": 193
+    },
+    {
+      "epoch": 0.8151260504201681,
+      "grad_norm": 4.740433606210983,
+      "learning_rate": 1.5068493150684933e-05,
+      "loss": 1.9741,
+      "step": 194
+    },
+    {
+      "epoch": 0.819327731092437,
+      "grad_norm": 6.156696419240408,
+      "learning_rate": 1.4956521739130436e-05,
+      "loss": 2.0031,
+      "step": 195
+    },
+    {
+      "epoch": 0.8235294117647058,
+      "grad_norm": 5.245506871359339,
+      "learning_rate": 1.4840989399293289e-05,
+      "loss": 1.9187,
+      "step": 196
+    },
+    {
+      "epoch": 0.8277310924369747,
+      "grad_norm": 3.76808185910061,
+      "learning_rate": 1.4721723518850987e-05,
+      "loss": 1.9746,
+      "step": 197
+    },
+    {
+      "epoch": 0.8319327731092437,
+      "grad_norm": 3.927325479381158,
+      "learning_rate": 1.45985401459854e-05,
+      "loss": 2.0478,
+      "step": 198
+    },
+    {
+      "epoch": 0.8361344537815126,
+      "grad_norm": 4.793479386506119,
+      "learning_rate": 1.4471243042671615e-05,
+      "loss": 1.9702,
+      "step": 199
+    },
+    {
+      "epoch": 0.8403361344537815,
+      "grad_norm": 3.4378170935260135,
+      "learning_rate": 1.4339622641509435e-05,
+      "loss": 1.9936,
+      "step": 200
+    },
+    {
+      "epoch": 0.8445378151260504,
+      "grad_norm": 5.645999605577803,
+      "learning_rate": 1.4203454894433781e-05,
+      "loss": 2.0336,
+      "step": 201
+    },
+    {
+      "epoch": 0.8487394957983193,
+      "grad_norm": 5.317794463649019,
+      "learning_rate": 1.4062500000000001e-05,
+      "loss": 2.0087,
+      "step": 202
+    },
+    {
+      "epoch": 0.8529411764705882,
+      "grad_norm": 3.350662535496213,
+      "learning_rate": 1.3916500994035784e-05,
+      "loss": 1.9442,
+      "step": 203
+    },
+    {
+      "epoch": 0.8571428571428571,
+      "grad_norm": 2.8975061156814137,
+      "learning_rate": 1.3765182186234817e-05,
+      "loss": 1.9664,
+      "step": 204
+    },
+    {
+      "epoch": 0.8613445378151261,
+      "grad_norm": 4.912078857539673,
+      "learning_rate": 1.3608247422680415e-05,
+      "loss": 1.9891,
+      "step": 205
+    },
+    {
+      "epoch": 0.865546218487395,
+      "grad_norm": 3.8579017476693207,
+      "learning_rate": 1.3445378151260506e-05,
+      "loss": 1.9428,
+      "step": 206
+    },
+    {
+      "epoch": 0.8697478991596639,
+      "grad_norm": 5.922561630455319,
+      "learning_rate": 1.3276231263383297e-05,
+      "loss": 2.0086,
+      "step": 207
+    },
+    {
+      "epoch": 0.8739495798319328,
+      "grad_norm": 5.7848073006616545,
+      "learning_rate": 1.3100436681222708e-05,
+      "loss": 1.9327,
+      "step": 208
+    },
+    {
+      "epoch": 0.8781512605042017,
+      "grad_norm": 2.0024190162323516,
+      "learning_rate": 1.2917594654788418e-05,
+      "loss": 2.0117,
+      "step": 209
+    },
+    {
+      "epoch": 0.8823529411764706,
+      "grad_norm": 3.0320546138824223,
+      "learning_rate": 1.2727272727272728e-05,
+      "loss": 1.9501,
+      "step": 210
+    },
+    {
+      "epoch": 0.8865546218487395,
+      "grad_norm": 3.8228323091811984,
+      "learning_rate": 1.2529002320185617e-05,
+      "loss": 1.9509,
+      "step": 211
+    },
+    {
+      "epoch": 0.8907563025210085,
+      "grad_norm": 3.1582266115572453,
+      "learning_rate": 1.2322274881516586e-05,
+      "loss": 2.0187,
+      "step": 212
+    },
+    {
+      "epoch": 0.8949579831932774,
+      "grad_norm": 3.7850080014396923,
+      "learning_rate": 1.2106537530266345e-05,
+      "loss": 1.9455,
+      "step": 213
+    },
+    {
+      "epoch": 0.8991596638655462,
+      "grad_norm": 3.055691850405243,
+      "learning_rate": 1.1881188118811881e-05,
+      "loss": 1.9306,
+      "step": 214
+    },
+    {
+      "epoch": 0.9033613445378151,
+      "grad_norm": 3.6195376384441404,
+      "learning_rate": 1.1645569620253165e-05,
+      "loss": 1.9476,
+      "step": 215
+    },
+    {
+      "epoch": 0.907563025210084,
+      "grad_norm": 2.792901728427093,
+      "learning_rate": 1.139896373056995e-05,
+      "loss": 1.9707,
+      "step": 216
+    },
+    {
+      "epoch": 0.9117647058823529,
+      "grad_norm": 8.809256342899616,
+      "learning_rate": 1.1140583554376659e-05,
+      "loss": 2.0588,
+      "step": 217
+    },
+    {
+      "epoch": 0.9159663865546218,
+      "grad_norm": 6.535409467311721,
+      "learning_rate": 1.0869565217391305e-05,
+      "loss": 1.9307,
+      "step": 218
+    },
+    {
+      "epoch": 0.9201680672268907,
+      "grad_norm": 3.6348451072738808,
+      "learning_rate": 1.0584958217270197e-05,
+      "loss": 1.9913,
+      "step": 219
+    },
+    {
+      "epoch": 0.9243697478991597,
+      "grad_norm": 5.190364501099717,
+      "learning_rate": 1.0285714285714285e-05,
+      "loss": 1.9926,
+      "step": 220
+    },
+    {
+      "epoch": 0.9285714285714286,
+      "grad_norm": 5.594530017203908,
+      "learning_rate": 9.970674486803521e-06,
+      "loss": 1.9809,
+      "step": 221
+    },
+    {
+      "epoch": 0.9327731092436975,
+      "grad_norm": 2.0779019236986174,
+      "learning_rate": 9.63855421686747e-06,
+      "loss": 1.9385,
+      "step": 222
+    },
+    {
+      "epoch": 0.9369747899159664,
+      "grad_norm": 4.796813646923566,
+      "learning_rate": 9.287925696594427e-06,
+      "loss": 2.0104,
+      "step": 223
+    },
+    {
+      "epoch": 0.9411764705882353,
+      "grad_norm": 4.778634494632563,
+      "learning_rate": 8.9171974522293e-06,
+      "loss": 1.9969,
+      "step": 224
+    },
+    {
+      "epoch": 0.9453781512605042,
+      "grad_norm": 2.8363001174562594,
+      "learning_rate": 8.524590163934427e-06,
+      "loss": 2.0084,
+      "step": 225
+    },
+    {
+      "epoch": 0.9495798319327731,
+      "grad_norm": 3.675032560462798,
+      "learning_rate": 8.108108108108107e-06,
+      "loss": 1.9197,
+      "step": 226
+    },
+    {
+      "epoch": 0.9537815126050421,
+      "grad_norm": 3.969674201409064,
+      "learning_rate": 7.665505226480837e-06,
+      "loss": 2.0004,
+      "step": 227
+    },
+    {
+      "epoch": 0.957983193277311,
+      "grad_norm": 2.4232749721231808,
+      "learning_rate": 7.194244604316547e-06,
+      "loss": 1.9574,
+      "step": 228
+    },
+    {
+      "epoch": 0.9621848739495799,
+      "grad_norm": 2.344162735674204,
+      "learning_rate": 6.691449814126395e-06,
+      "loss": 1.985,
+      "step": 229
+    },
+    {
+      "epoch": 0.9663865546218487,
+      "grad_norm": 3.0362841354288643,
+      "learning_rate": 6.153846153846153e-06,
+      "loss": 1.9752,
+      "step": 230
+    },
+    {
+      "epoch": 0.9705882352941176,
+      "grad_norm": 1.6752526393389675,
+      "learning_rate": 5.577689243027889e-06,
+      "loss": 1.9744,
+      "step": 231
+    },
+    {
+      "epoch": 0.9747899159663865,
+      "grad_norm": 2.076504652779221,
+      "learning_rate": 4.958677685950414e-06,
+      "loss": 1.9806,
+      "step": 232
+    },
+    {
+      "epoch": 0.9789915966386554,
+      "grad_norm": 5.232894588002972,
+      "learning_rate": 4.291845493562232e-06,
+      "loss": 1.947,
+      "step": 233
+    },
+    {
+      "epoch": 0.9831932773109243,
+      "grad_norm": 2.3873723202577177,
+      "learning_rate": 3.5714285714285714e-06,
+      "loss": 1.9534,
+      "step": 234
+    },
+    {
+      "epoch": 0.9873949579831933,
+      "grad_norm": 3.0394990303644245,
+      "learning_rate": 2.790697674418604e-06,
+      "loss": 1.9814,
+      "step": 235
+    },
+    {
+      "epoch": 0.9915966386554622,
+      "grad_norm": 1.662686705633431,
+      "learning_rate": 1.941747572815534e-06,
+      "loss": 1.9857,
+      "step": 236
+    },
+    {
+      "epoch": 0.9957983193277311,
+      "grad_norm": 1.2954300657585052,
+      "learning_rate": 1.0152284263959392e-06,
+      "loss": 1.9733,
+      "step": 237
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 5.354785357697052,
+      "learning_rate": 0,
+      "loss": 1.9467,
+      "step": 238
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 238,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 119,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.01932997936939e+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 sha256:82425111d1e98101eb231559b821826cb3342caa58a34cc1d8e12678d817143e
+size 9105

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)