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

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

config.json

@@ -0,0 +1,28 @@
+{
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.51.3",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,14 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.51.3"
+}

latest

@@ -0,0 +1 @@
+global_step198

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e0ad8a20391d004e08f908af0a931cc03bdfd2f17e7ed7ea1bafb4cb4e5bf80
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a1528ac61dee840904da08e7a3f991ce2bdf69bc6c536c9ae6c2d043cf84df6
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10a546f9ee785ca28310032ead75c5ec8996d28085167e63dc25ae2148b650a2
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50597525d6f42400c7541391256302bdfde8bf4d2ead4dcf4d76c5c355575211
+size 1089994880

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15231233024
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
+    "model.norm.weight": "model-00003-of-00004.safetensors"
+  }
+}

rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70208c3fb44e5fcccd73f2847648badd160b8f3bad7a5421ec65a5dbafba6df3
+size 14768

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7851155a86223a21c7a3354af9a64bb7de634603b85a0cccba1626c89121d4e3
+size 14768

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09e831fd61bd5abc7cf91627888b51bdfb3232dc5b5ca24beab591a854d5d401
+size 14768

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:603067db864b1ab6d8c72f3ade56691eea647663c3c1aff0f89036d121586635
+size 1064

special_tokens_map.json

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

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5eee858c5123a4279c3e1f7b81247343f356ac767940b2692a928ad929543214
+size 11422063

tokenizer_config.json

@@ -0,0 +1,209 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,1840 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.998109640831758,
+  "eval_steps": 51,
+  "global_step": 198,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.15104166666666666,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 975.3333333333334,
+      "completions/mean_length": 312.0329996744792,
+      "completions/mean_terminated_length": 185.89303080240884,
+      "completions/min_length": 28.333333333333332,
+      "completions/min_terminated_length": 28.333333333333332,
+      "epoch": 0.045368620037807186,
+      "grad_norm": 0.14972379803657532,
+      "kl": 4.560748736063639e-05,
+      "learning_rate": 4e-07,
+      "loss": -0.0081,
+      "num_tokens": 942182.0,
+      "reward": 0.37008477250734967,
+      "reward_std": 0.11998833467562993,
+      "rewards/get_embedding_sim/mean": 0.3440430959065755,
+      "rewards/get_embedding_sim/std": 0.06710867583751678,
+      "rewards/reward_num_unique_chars/mean": 0.026041666666666668,
+      "rewards/reward_num_unique_chars/std": 0.14761295169591904,
+      "step": 3
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.13020833333333334,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 959.0,
+      "completions/mean_length": 307.0069580078125,
+      "completions/mean_terminated_length": 199.09521484375,
+      "completions/min_length": 10.333333333333334,
+      "completions/min_terminated_length": 10.333333333333334,
+      "epoch": 0.09073724007561437,
+      "grad_norm": 0.12008437514305115,
+      "kl": 0.0001388813058535258,
+      "learning_rate": 1e-06,
+      "loss": 0.035,
+      "num_tokens": 1882942.0,
+      "reward": 0.4796616733074188,
+      "reward_std": 0.214401513338089,
+      "rewards/get_embedding_sim/mean": 0.3694185713926951,
+      "rewards/get_embedding_sim/std": 0.07585694640874863,
+      "rewards/reward_num_unique_chars/mean": 0.1102430559694767,
+      "rewards/reward_num_unique_chars/std": 0.2982482860485713,
+      "step": 6
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.08072916666666667,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 845.0,
+      "completions/mean_length": 234.67969258626303,
+      "completions/mean_terminated_length": 166.36500040690103,
+      "completions/min_length": 21.0,
+      "completions/min_terminated_length": 21.0,
+      "epoch": 0.13610586011342155,
+      "grad_norm": 0.08606597781181335,
+      "kl": 0.00013801626240213713,
+      "learning_rate": 1e-06,
+      "loss": 0.0172,
+      "num_tokens": 2735293.0,
+      "reward": 0.39071526130040485,
+      "reward_std": 0.1662569542725881,
+      "rewards/get_embedding_sim/mean": 0.33168746034304303,
+      "rewards/get_embedding_sim/std": 0.07500659177700679,
+      "rewards/reward_num_unique_chars/mean": 0.059027779226501785,
+      "rewards/reward_num_unique_chars/std": 0.22509141763051352,
+      "step": 9
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.10503472222222225,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 983.3333333333334,
+      "completions/mean_length": 262.79688517252606,
+      "completions/mean_terminated_length": 173.54302469889322,
+      "completions/min_length": 12.0,
+      "completions/min_terminated_length": 12.0,
+      "epoch": 0.18147448015122875,
+      "grad_norm": 0.11949238181114197,
+      "kl": 0.00030877192815144855,
+      "learning_rate": 1e-06,
+      "loss": -0.0001,
+      "num_tokens": 3627859.0,
+      "reward": 0.4095470607280731,
+      "reward_std": 0.18979967882235846,
+      "rewards/get_embedding_sim/mean": 0.33055397868156433,
+      "rewards/get_embedding_sim/std": 0.07462155818939209,
+      "rewards/reward_num_unique_chars/mean": 0.07899305472771327,
+      "rewards/reward_num_unique_chars/std": 0.25569593409697217,
+      "step": 12
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.1362847222222222,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 991.6666666666666,
+      "completions/mean_length": 316.6762288411458,
+      "completions/mean_terminated_length": 204.85944112141928,
+      "completions/min_length": 9.666666666666666,
+      "completions/min_terminated_length": 9.666666666666666,
+      "epoch": 0.22684310018903592,
+      "grad_norm": 0.16435399651527405,
+      "kl": 0.0005876521269480387,
+      "learning_rate": 1e-06,
+      "loss": 0.0529,
+      "num_tokens": 4554894.0,
+      "reward": 0.4522427221139272,
+      "reward_std": 0.205996572971344,
+      "rewards/get_embedding_sim/mean": 0.35502047340075177,
+      "rewards/get_embedding_sim/std": 0.076506607234478,
+      "rewards/reward_num_unique_chars/mean": 0.09722222139437993,
+      "rewards/reward_num_unique_chars/std": 0.27809616923332214,
+      "step": 15
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.11718750000000004,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 927.0,
+      "completions/mean_length": 283.77171834309894,
+      "completions/mean_terminated_length": 184.45149739583334,
+      "completions/min_length": 24.666666666666668,
+      "completions/min_terminated_length": 24.666666666666668,
+      "epoch": 0.2722117202268431,
+      "grad_norm": 0.17904439568519592,
+      "kl": 0.0004306634267171224,
+      "learning_rate": 1e-06,
+      "loss": 0.036,
+      "num_tokens": 5464567.0,
+      "reward": 0.47324784596761066,
+      "reward_std": 0.2480545292297999,
+      "rewards/get_embedding_sim/mean": 0.35345616936683655,
+      "rewards/get_embedding_sim/std": 0.08570993691682816,
+      "rewards/reward_num_unique_chars/mean": 0.11979166915019353,
+      "rewards/reward_num_unique_chars/std": 0.32309961318969727,
+      "step": 18
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.07204861111111112,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 1011.0,
+      "completions/mean_length": 230.54688517252603,
+      "completions/mean_terminated_length": 169.21332804361978,
+      "completions/min_length": 12.333333333333334,
+      "completions/min_terminated_length": 12.333333333333334,
+      "epoch": 0.31758034026465026,
+      "grad_norm": 0.11123450100421906,
+      "kl": 0.0011239051818847656,
+      "learning_rate": 1e-06,
+      "loss": 0.0222,
+      "num_tokens": 6313117.0,
+      "reward": 0.4715224802494049,
+      "reward_std": 0.2366275986035665,
+      "rewards/get_embedding_sim/mean": 0.3491266171137492,
+      "rewards/get_embedding_sim/std": 0.06465367351969083,
+      "rewards/reward_num_unique_chars/mean": 0.1223958358168602,
+      "rewards/reward_num_unique_chars/std": 0.3250391185283661,
+      "step": 21
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.09895833333333337,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 961.0,
+      "completions/mean_length": 229.36719258626303,
+      "completions/mean_terminated_length": 142.55723571777344,
+      "completions/min_length": 15.0,
+      "completions/min_terminated_length": 15.0,
+      "epoch": 0.3629489603024575,
+      "grad_norm": 0.118320994079113,
+      "kl": 0.0019257068634033203,
+      "learning_rate": 1e-06,
+      "loss": 0.0158,
+      "num_tokens": 7162132.0,
+      "reward": 0.5189645787080129,
+      "reward_std": 0.24159842729568481,
+      "rewards/get_embedding_sim/mean": 0.3809437155723572,
+      "rewards/get_embedding_sim/std": 0.0799456536769867,
+      "rewards/reward_num_unique_chars/mean": 0.13802083084980646,
+      "rewards/reward_num_unique_chars/std": 0.3419287900129954,
+      "step": 24
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.052951388888888874,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 982.3333333333334,
+      "completions/mean_length": 197.44445292154947,
+      "completions/mean_terminated_length": 151.25631205240884,
+      "completions/min_length": 14.666666666666666,
+      "completions/min_terminated_length": 14.666666666666666,
+      "epoch": 0.40831758034026466,
+      "grad_norm": 0.11851406842470169,
+      "kl": 0.002936681111653646,
+      "learning_rate": 1e-06,
+      "loss": 0.0317,
+      "num_tokens": 7973172.0,
+      "reward": 0.569815476735433,
+      "reward_std": 0.25512967507044476,
+      "rewards/get_embedding_sim/mean": 0.362350195646286,
+      "rewards/get_embedding_sim/std": 0.07909337679545085,
+      "rewards/reward_num_unique_chars/mean": 0.2074652761220932,
+      "rewards/reward_num_unique_chars/std": 0.4044720729192098,
+      "step": 27
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.056423611111111126,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 934.3333333333334,
+      "completions/mean_length": 211.5963592529297,
+      "completions/mean_terminated_length": 163.11248270670572,
+      "completions/min_length": 16.333333333333332,
+      "completions/min_terminated_length": 16.333333333333332,
+      "epoch": 0.45368620037807184,
+      "grad_norm": 0.21573348343372345,
+      "kl": 0.008742332458496094,
+      "learning_rate": 1e-06,
+      "loss": 0.0125,
+      "num_tokens": 8794371.0,
+      "reward": 0.43826034665107727,
+      "reward_std": 0.20837691922982535,
+      "rewards/get_embedding_sim/mean": 0.3427741924921672,
+      "rewards/get_embedding_sim/std": 0.0719177375237147,
+      "rewards/reward_num_unique_chars/mean": 0.09548610945542653,
+      "rewards/reward_num_unique_chars/std": 0.2681623448928197,
+      "step": 30
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.056423611111111126,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 892.0,
+      "completions/mean_length": 197.90365091959634,
+      "completions/mean_terminated_length": 148.50442504882812,
+      "completions/min_length": 15.0,
+      "completions/min_terminated_length": 15.0,
+      "epoch": 0.499054820415879,
+      "grad_norm": 0.08199404180049896,
+      "kl": 0.005775133768717448,
+      "learning_rate": 1e-06,
+      "loss": 0.0113,
+      "num_tokens": 9601812.0,
+      "reward": 0.45480871200561523,
+      "reward_std": 0.2194500764211019,
+      "rewards/get_embedding_sim/mean": 0.36192673444747925,
+      "rewards/get_embedding_sim/std": 0.0750991627573967,
+      "rewards/reward_num_unique_chars/mean": 0.0928819440305233,
+      "rewards/reward_num_unique_chars/std": 0.2857237259546916,
+      "step": 33
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.053819444444444454,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 873.3333333333334,
+      "completions/mean_length": 201.65104166666666,
+      "completions/mean_terminated_length": 154.5564727783203,
+      "completions/min_length": 10.666666666666666,
+      "completions/min_terminated_length": 10.666666666666666,
+      "epoch": 0.5444234404536862,
+      "grad_norm": 0.13542793691158295,
+      "kl": 0.011366526285807291,
+      "learning_rate": 1e-06,
+      "loss": -0.0008,
+      "num_tokens": 10414722.0,
+      "reward": 0.4134095311164856,
+      "reward_std": 0.16343241184949875,
+      "rewards/get_embedding_sim/mean": 0.3708748022715251,
+      "rewards/get_embedding_sim/std": 0.08833041042089462,
+      "rewards/reward_num_unique_chars/mean": 0.042534722636143364,
+      "rewards/reward_num_unique_chars/std": 0.1979833443959554,
+      "step": 36
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.04340277777777779,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 891.6666666666666,
+      "completions/mean_length": 176.1076456705729,
+      "completions/mean_terminated_length": 137.35225423177084,
+      "completions/min_length": 11.0,
+      "completions/min_terminated_length": 11.0,
+      "epoch": 0.5897920604914934,
+      "grad_norm": 1.7642544507980347,
+      "kl": 0.151151974995931,
+      "learning_rate": 1e-06,
+      "loss": 0.0179,
+      "num_tokens": 11207422.0,
+      "reward": 0.5713514387607574,
+      "reward_std": 0.26335498690605164,
+      "rewards/get_embedding_sim/mean": 0.36909447113672894,
+      "rewards/get_embedding_sim/std": 0.09187572946151097,
+      "rewards/reward_num_unique_chars/mean": 0.202256940305233,
+      "rewards/reward_num_unique_chars/std": 0.390445997317632,
+      "step": 39
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.032986111111111126,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 826.3333333333334,
+      "completions/mean_length": 201.41754150390625,
+      "completions/mean_terminated_length": 173.10018412272134,
+      "completions/min_length": 18.333333333333332,
+      "completions/min_terminated_length": 18.333333333333332,
+      "epoch": 0.6351606805293005,
+      "grad_norm": 0.10251538455486298,
+      "kl": 0.014621734619140625,
+      "learning_rate": 1e-06,
+      "loss": 0.0102,
+      "num_tokens": 12029279.0,
+      "reward": 0.5142592787742615,
+      "reward_std": 0.2620675365130107,
+      "rewards/get_embedding_sim/mean": 0.3701620002587636,
+      "rewards/get_embedding_sim/std": 0.10092929750680923,
+      "rewards/reward_num_unique_chars/mean": 0.1440972238779068,
+      "rewards/reward_num_unique_chars/std": 0.34582529465357464,
+      "step": 42
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.026041666666666668,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 927.6666666666666,
+      "completions/mean_length": 183.04254150390625,
+      "completions/mean_terminated_length": 160.71800740559897,
+      "completions/min_length": 14.666666666666666,
+      "completions/min_terminated_length": 14.666666666666666,
+      "epoch": 0.6805293005671077,
+      "grad_norm": 0.09084329754114151,
+      "kl": 0.015349706013997396,
+      "learning_rate": 1e-06,
+      "loss": -0.0004,
+      "num_tokens": 12816000.0,
+      "reward": 0.5384640991687775,
+      "reward_std": 0.22944432497024536,
+      "rewards/get_embedding_sim/mean": 0.39697099725405377,
+      "rewards/get_embedding_sim/std": 0.10396929830312729,
+      "rewards/reward_num_unique_chars/mean": 0.14149305721124014,
+      "rewards/reward_num_unique_chars/std": 0.3254843403895696,
+      "step": 45
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.02777777777777779,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 802.3333333333334,
+      "completions/mean_length": 165.0295155843099,
+      "completions/mean_terminated_length": 140.43072509765625,
+      "completions/min_length": 9.666666666666666,
+      "completions/min_terminated_length": 9.666666666666666,
+      "epoch": 0.725897920604915,
+      "grad_norm": 0.21910759806632996,
+      "kl": 0.027149200439453125,
+      "learning_rate": 1e-06,
+      "loss": 0.0009,
+      "num_tokens": 13587394.0,
+      "reward": 0.5553397635618845,
+      "reward_std": 0.23784717917442322,
+      "rewards/get_embedding_sim/mean": 0.4086383481820424,
+      "rewards/get_embedding_sim/std": 0.10949051380157471,
+      "rewards/reward_num_unique_chars/mean": 0.14670138930281004,
+      "rewards/reward_num_unique_chars/std": 0.33698558807373047,
+      "step": 48
+    },
+    {
+      "epoch": 0.7712665406427222,
+      "grad_norm": 0.09893961995840073,
+      "learning_rate": 1e-06,
+      "loss": 0.0047,
+      "step": 51
+    },
+    {
+      "epoch": 0.7712665406427222,
+      "eval_clip_ratio/high_max": 0.0,
+      "eval_clip_ratio/high_mean": 0.0,
+      "eval_clip_ratio/low_mean": 0.0,
+      "eval_clip_ratio/low_min": 0.0,
+      "eval_clip_ratio/region_mean": 0.0,
+      "eval_completions/clipped_ratio": 0.12797619047619044,
+      "eval_completions/max_length": 880.7678571428571,
+      "eval_completions/max_terminated_length": 701.3214285714286,
+      "eval_completions/mean_length": 258.24070589882984,
+      "eval_completions/mean_terminated_length": 153.6624070576259,
+      "eval_completions/min_length": 24.446428571428573,
+      "eval_completions/min_terminated_length": 24.446428571428573,
+      "eval_kl": 0.0542449951171875,
+      "eval_loss": 0.026244351640343666,
+      "eval_num_tokens": 14351398.0,
+      "eval_reward": 0.524820977555854,
+      "eval_reward_std": 0.22432494928528154,
+      "eval_rewards/get_embedding_sim/mean": 0.43479119294456076,
+      "eval_rewards/get_embedding_sim/std": 0.09110667330345937,
+      "eval_rewards/reward_num_unique_chars/mean": 0.09002976235933602,
+      "eval_rewards/reward_num_unique_chars/std": 0.18600706889161042,
+      "eval_runtime": 2254.2404,
+      "eval_samples_per_second": 0.025,
+      "eval_steps_per_second": 0.001,
+      "step": 51
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.044270833333333315,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 921.3333333333334,
+      "completions/mean_length": 197.77517954508463,
+      "completions/mean_terminated_length": 159.97277196248373,
+      "completions/min_length": 9.833333333333334,
+      "completions/min_terminated_length": 9.833333333333334,
+      "epoch": 0.8166351606805293,
+      "grad_norm": 0.08635270595550537,
+      "kl": 0.030397415161132812,
+      "learning_rate": 1e-06,
+      "loss": 0.0077,
+      "num_tokens": 15200636.0,
+      "reward": 0.5215439548095068,
+      "reward_std": 0.23126975446939468,
+      "rewards/get_embedding_sim/mean": 0.42692585786183673,
+      "rewards/get_embedding_sim/std": 0.11467409133911133,
+      "rewards/reward_num_unique_chars/mean": 0.09461805845300357,
+      "rewards/reward_num_unique_chars/std": 0.28477593511343,
+      "step": 54
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.04253472222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 836.6666666666666,
+      "completions/mean_length": 198.4244842529297,
+      "completions/mean_terminated_length": 161.79749043782553,
+      "completions/min_length": 10.333333333333334,
+      "completions/min_terminated_length": 10.333333333333334,
+      "epoch": 0.8620037807183365,
+      "grad_norm": 14.726771354675293,
+      "kl": 0.21588261922200522,
+      "learning_rate": 1e-06,
+      "loss": 0.0104,
+      "num_tokens": 16019045.0,
+      "reward": 0.5494122306505839,
+      "reward_std": 0.24494746327400208,
+      "rewards/get_embedding_sim/mean": 0.44264134764671326,
+      "rewards/get_embedding_sim/std": 0.11085022240877151,
+      "rewards/reward_num_unique_chars/mean": 0.10677083333333333,
+      "rewards/reward_num_unique_chars/std": 0.30227985978126526,
+      "step": 57
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.04253472222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 964.6666666666666,
+      "completions/mean_length": 200.72309366861978,
+      "completions/mean_terminated_length": 164.248779296875,
+      "completions/min_length": 11.333333333333334,
+      "completions/min_terminated_length": 11.333333333333334,
+      "epoch": 0.9073724007561437,
+      "grad_norm": 0.09581304341554642,
+      "kl": 0.33023325602213544,
+      "learning_rate": 1e-06,
+      "loss": 0.0072,
+      "num_tokens": 16832758.0,
+      "reward": 0.599389910697937,
+      "reward_std": 0.26327316959698993,
+      "rewards/get_embedding_sim/mean": 0.45268850525220233,
+      "rewards/get_embedding_sim/std": 0.11441038797299068,
+      "rewards/reward_num_unique_chars/mean": 0.14670139302810034,
+      "rewards/reward_num_unique_chars/std": 0.31440146267414093,
+      "step": 60
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.021701388888888878,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 917.3333333333334,
+      "completions/mean_length": 186.21094258626303,
+      "completions/mean_terminated_length": 167.5730946858724,
+      "completions/min_length": 13.0,
+      "completions/min_terminated_length": 13.0,
+      "epoch": 0.9527410207939508,
+      "grad_norm": 0.08248484879732132,
+      "kl": 0.04541015625,
+      "learning_rate": 1e-06,
+      "loss": 0.0017,
+      "num_tokens": 17637097.0,
+      "reward": 0.5855847001075745,
+      "reward_std": 0.2750825683275859,
+      "rewards/get_embedding_sim/mean": 0.46405691901842755,
+      "rewards/get_embedding_sim/std": 0.11442819982767105,
+      "rewards/reward_num_unique_chars/mean": 0.12152778108914693,
+      "rewards/reward_num_unique_chars/std": 0.3193853000799815,
+      "step": 63
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.031507423371647504,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 930.0,
+      "completions/mean_length": 193.8086140950521,
+      "completions/mean_terminated_length": 167.16290283203125,
+      "completions/min_length": 7.0,
+      "completions/min_terminated_length": 7.0,
+      "epoch": 0.998109640831758,
+      "grad_norm": 0.06374574452638626,
+      "kl": 0.03699493408203125,
+      "learning_rate": 1e-06,
+      "loss": 0.0187,
+      "num_tokens": 18440914.0,
+      "reward": 0.6297420461972555,
+      "reward_std": 0.2834969659646352,
+      "rewards/get_embedding_sim/mean": 0.47088783979415894,
+      "rewards/get_embedding_sim/std": 0.11324869592984517,
+      "rewards/reward_num_unique_chars/mean": 0.1588541641831398,
+      "rewards/reward_num_unique_chars/std": 0.36384791135787964,
+      "step": 66
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.032118055555555546,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 833.3333333333334,
+      "completions/mean_length": 200.6024373372396,
+      "completions/mean_terminated_length": 173.2165069580078,
+      "completions/min_length": 9.0,
+      "completions/min_terminated_length": 9.0,
+      "epoch": 1.0453686200378072,
+      "grad_norm": 0.11849670857191086,
+      "kl": 0.05316925048828125,
+      "learning_rate": 1e-06,
+      "loss": 0.0071,
+      "num_tokens": 19261832.0,
+      "reward": 0.5802033940951029,
+      "reward_std": 0.25838569800059,
+      "rewards/get_embedding_sim/mean": 0.4734325309594472,
+      "rewards/get_embedding_sim/std": 0.11253533015648524,
+      "rewards/reward_num_unique_chars/mean": 0.10677083084980647,
+      "rewards/reward_num_unique_chars/std": 0.30244183043638867,
+      "step": 69
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.039930555555555546,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 973.0,
+      "completions/mean_length": 199.55555725097656,
+      "completions/mean_terminated_length": 165.2902577718099,
+      "completions/min_length": 9.333333333333334,
+      "completions/min_terminated_length": 9.333333333333334,
+      "epoch": 1.0907372400756143,
+      "grad_norm": 0.10332732647657394,
+      "kl": 0.0515289306640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0075,
+      "num_tokens": 20067096.0,
+      "reward": 0.625789741675059,
+      "reward_std": 0.2765499949455261,
+      "rewards/get_embedding_sim/mean": 0.49471331636110943,
+      "rewards/get_embedding_sim/std": 0.11266019940376282,
+      "rewards/reward_num_unique_chars/mean": 0.1310763880610466,
+      "rewards/reward_num_unique_chars/std": 0.336679349342982,
+      "step": 72
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.036458333333333294,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 1008.6666666666666,
+      "completions/mean_length": 209.8359375,
+      "completions/mean_terminated_length": 178.97360229492188,
+      "completions/min_length": 7.0,
+      "completions/min_terminated_length": 7.0,
+      "epoch": 1.1361058601134215,
+      "grad_norm": 0.11925654858350754,
+      "kl": 0.14461263020833334,
+      "learning_rate": 1e-06,
+      "loss": 0.0055,
+      "num_tokens": 20893467.0,
+      "reward": 0.5831413467725118,
+      "reward_std": 0.2582869480053584,
+      "rewards/get_embedding_sim/mean": 0.4919954836368561,
+      "rewards/get_embedding_sim/std": 0.1114387462536494,
+      "rewards/reward_num_unique_chars/mean": 0.09114583333333333,
+      "rewards/reward_num_unique_chars/std": 0.2839343051115672,
+      "step": 75
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.047743055555555546,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 897.3333333333334,
+      "completions/mean_length": 221.0104217529297,
+      "completions/mean_terminated_length": 180.71256510416666,
+      "completions/min_length": 5.0,
+      "completions/min_terminated_length": 5.0,
+      "epoch": 1.1814744801512287,
+      "grad_norm": 0.09554021060466766,
+      "kl": 0.16336822509765625,
+      "learning_rate": 1e-06,
+      "loss": 0.0197,
+      "num_tokens": 21730887.0,
+      "reward": 0.6385945876439413,
+      "reward_std": 0.2661168724298477,
+      "rewards/get_embedding_sim/mean": 0.5127264857292175,
+      "rewards/get_embedding_sim/std": 0.11183823893467586,
+      "rewards/reward_num_unique_chars/mean": 0.12586805472771326,
+      "rewards/reward_num_unique_chars/std": 0.3207412262757619,
+      "step": 78
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.049479166666666664,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 939.0,
+      "completions/mean_length": 228.04601033528647,
+      "completions/mean_terminated_length": 186.1550038655599,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.2268431001890359,
+      "grad_norm": 0.07755686342716217,
+      "kl": 0.05751800537109375,
+      "learning_rate": 1e-06,
+      "loss": 0.0156,
+      "num_tokens": 22583420.0,
+      "reward": 0.6019672354062399,
+      "reward_std": 0.26383428772290546,
+      "rewards/get_embedding_sim/mean": 0.5021408100922903,
+      "rewards/get_embedding_sim/std": 0.10627821832895279,
+      "rewards/reward_num_unique_chars/mean": 0.09982638930281003,
+      "rewards/reward_num_unique_chars/std": 0.2837299009164174,
+      "step": 81
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.029513888888888878,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 931.3333333333334,
+      "completions/mean_length": 191.74740091959634,
+      "completions/mean_terminated_length": 166.4415028889974,
+      "completions/min_length": 8.666666666666666,
+      "completions/min_terminated_length": 8.666666666666666,
+      "epoch": 1.272211720226843,
+      "grad_norm": 0.08697984367609024,
+      "kl": 0.057329813639322914,
+      "learning_rate": 1e-06,
+      "loss": 0.0141,
+      "num_tokens": 23394137.0,
+      "reward": 0.6638144056002299,
+      "reward_std": 0.26522762576738995,
+      "rewards/get_embedding_sim/mean": 0.5231893658638,
+      "rewards/get_embedding_sim/std": 0.10482257604598999,
+      "rewards/reward_num_unique_chars/mean": 0.140625,
+      "rewards/reward_num_unique_chars/std": 0.3480878472328186,
+      "step": 84
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.026041666666666668,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 872.3333333333334,
+      "completions/mean_length": 185.42535400390625,
+      "completions/mean_terminated_length": 163.15696716308594,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.3175803402646502,
+      "grad_norm": 0.14970338344573975,
+      "kl": 0.12465922037760417,
+      "learning_rate": 1e-06,
+      "loss": 0.008,
+      "num_tokens": 24197571.0,
+      "reward": 0.6190575559933981,
+      "reward_std": 0.2601381540298462,
+      "rewards/get_embedding_sim/mean": 0.5174950361251831,
+      "rewards/get_embedding_sim/std": 0.0997606838742892,
+      "rewards/reward_num_unique_chars/mean": 0.10156250124176343,
+      "rewards/reward_num_unique_chars/std": 0.297150323788325,
+      "step": 87
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.029513888888888912,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 965.6666666666666,
+      "completions/mean_length": 216.75694783528647,
+      "completions/mean_terminated_length": 192.3217315673828,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.3629489603024574,
+      "grad_norm": 0.11725780367851257,
+      "kl": 0.08345540364583333,
+      "learning_rate": 1e-06,
+      "loss": 0.0079,
+      "num_tokens": 25034555.0,
+      "reward": 0.5995156168937683,
+      "reward_std": 0.22840352356433868,
+      "rewards/get_embedding_sim/mean": 0.5118419329325358,
+      "rewards/get_embedding_sim/std": 0.0987908939520518,
+      "rewards/reward_num_unique_chars/mean": 0.0876736119389534,
+      "rewards/reward_num_unique_chars/std": 0.27722589671611786,
+      "step": 90
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.029513888888888878,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 884.3333333333334,
+      "completions/mean_length": 186.4375,
+      "completions/mean_terminated_length": 161.33899434407553,
+      "completions/min_length": 6.0,
+      "completions/min_terminated_length": 6.0,
+      "epoch": 1.4083175803402646,
+      "grad_norm": 0.10394510626792908,
+      "kl": 0.07037099202473958,
+      "learning_rate": 1e-06,
+      "loss": 0.0084,
+      "num_tokens": 25831283.0,
+      "reward": 0.6795124411582947,
+      "reward_std": 0.29141750435034436,
+      "rewards/get_embedding_sim/mean": 0.5137137969334921,
+      "rewards/get_embedding_sim/std": 0.09767910589774449,
+      "rewards/reward_num_unique_chars/mean": 0.16579860697189966,
+      "rewards/reward_num_unique_chars/std": 0.3542452355225881,
+      "step": 93
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.05034722222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 971.3333333333334,
+      "completions/mean_length": 212.23785400390625,
+      "completions/mean_terminated_length": 169.16080729166666,
+      "completions/min_length": 6.666666666666667,
+      "completions/min_terminated_length": 6.666666666666667,
+      "epoch": 1.4536862003780717,
+      "grad_norm": 0.10010381788015366,
+      "kl": 0.06960042317708333,
+      "learning_rate": 1e-06,
+      "loss": 0.0203,
+      "num_tokens": 26656485.0,
+      "reward": 0.6300086975097656,
+      "reward_std": 0.24619843065738678,
+      "rewards/get_embedding_sim/mean": 0.5267100731531779,
+      "rewards/get_embedding_sim/std": 0.1071697548031807,
+      "rewards/reward_num_unique_chars/mean": 0.1032986119389534,
+      "rewards/reward_num_unique_chars/std": 0.29826584458351135,
+      "step": 96
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.037326388888888874,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 919.0,
+      "completions/mean_length": 211.8107655843099,
+      "completions/mean_terminated_length": 180.1980183919271,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 1.499054820415879,
+      "grad_norm": 0.07485458254814148,
+      "kl": 0.061063130696614586,
+      "learning_rate": 1e-06,
+      "loss": 0.0248,
+      "num_tokens": 27490315.0,
+      "reward": 0.620047926902771,
+      "reward_std": 0.2632503807544708,
+      "rewards/get_embedding_sim/mean": 0.5132770538330078,
+      "rewards/get_embedding_sim/std": 0.10026986648639043,
+      "rewards/reward_num_unique_chars/mean": 0.10677083333333333,
+      "rewards/reward_num_unique_chars/std": 0.3043619990348816,
+      "step": 99
+    },
+    {
+      "epoch": 1.544423440453686,
+      "grad_norm": 0.11106861382722855,
+      "learning_rate": 1e-06,
+      "loss": 0.0115,
+      "step": 102
+    },
+    {
+      "epoch": 1.544423440453686,
+      "eval_clip_ratio/high_max": 0.0,
+      "eval_clip_ratio/high_mean": 0.0,
+      "eval_clip_ratio/low_mean": 0.0,
+      "eval_clip_ratio/low_min": 0.0,
+      "eval_clip_ratio/region_mean": 0.0,
+      "eval_completions/clipped_ratio": 0.07068452380952381,
+      "eval_completions/max_length": 887.5892857142857,
+      "eval_completions/max_terminated_length": 675.0892857142857,
+      "eval_completions/mean_length": 206.3244113922119,
+      "eval_completions/mean_terminated_length": 145.37539066587175,
+      "eval_completions/min_length": 18.160714285714285,
+      "eval_completions/min_terminated_length": 18.160714285714285,
+      "eval_kl": 0.06965419224330358,
+      "eval_loss": 0.03773626312613487,
+      "eval_num_tokens": 28307736.0,
+      "eval_reward": 0.6229457370936871,
+      "eval_reward_std": 0.2839882879384926,
+      "eval_rewards/get_embedding_sim/mean": 0.5206391582531589,
+      "eval_rewards/get_embedding_sim/std": 0.09148550758670483,
+      "eval_rewards/reward_num_unique_chars/mean": 0.10230654794057566,
+      "eval_rewards/reward_num_unique_chars/std": 0.24572753932859218,
+      "eval_runtime": 1726.6979,
+      "eval_samples_per_second": 0.032,
+      "eval_steps_per_second": 0.001,
+      "step": 102
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03602430555555556,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 990.3333333333334,
+      "completions/mean_length": 214.1545181274414,
+      "completions/mean_terminated_length": 183.94319661458334,
+      "completions/min_length": 6.5,
+      "completions/min_terminated_length": 6.5,
+      "epoch": 1.5897920604914932,
+      "grad_norm": 1.0838171243667603,
+      "kl": 0.0672899881998698,
+      "learning_rate": 1e-06,
+      "loss": 0.0253,
+      "num_tokens": 29138511.0,
+      "reward": 0.6624543964862823,
+      "reward_std": 0.26948046932617825,
+      "rewards/get_embedding_sim/mean": 0.5296418766180674,
+      "rewards/get_embedding_sim/std": 0.10213356713453929,
+      "rewards/reward_num_unique_chars/mean": 0.1328124993791183,
+      "rewards/reward_num_unique_chars/std": 0.32840434461832047,
+      "step": 105
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.026041666666666668,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 974.0,
+      "completions/mean_length": 201.72048950195312,
+      "completions/mean_terminated_length": 179.72496032714844,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.6351606805293004,
+      "grad_norm": 0.0918864831328392,
+      "kl": 0.061335245768229164,
+      "learning_rate": 1e-06,
+      "loss": 0.0208,
+      "num_tokens": 29960717.0,
+      "reward": 0.6120087305704752,
+      "reward_std": 0.250284880399704,
+      "rewards/get_embedding_sim/mean": 0.5364878376324972,
+      "rewards/get_embedding_sim/std": 0.0979540745417277,
+      "rewards/reward_num_unique_chars/mean": 0.07552083457509677,
+      "rewards/reward_num_unique_chars/std": 0.26320414741834003,
+      "step": 108
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03559027777777779,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 923.0,
+      "completions/mean_length": 206.97309366861978,
+      "completions/mean_terminated_length": 177.1915079752604,
+      "completions/min_length": 7.666666666666667,
+      "completions/min_terminated_length": 7.666666666666667,
+      "epoch": 1.6805293005671076,
+      "grad_norm": 0.07678642123937607,
+      "kl": 0.06285349527994792,
+      "learning_rate": 1e-06,
+      "loss": 0.0214,
+      "num_tokens": 30781870.0,
+      "reward": 0.6274827718734741,
+      "reward_std": 0.26556732257207233,
+      "rewards/get_embedding_sim/mean": 0.5155035257339478,
+      "rewards/get_embedding_sim/std": 0.09278701990842819,
+      "rewards/reward_num_unique_chars/mean": 0.11197916666666667,
+      "rewards/reward_num_unique_chars/std": 0.30655037860075635,
+      "step": 111
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.032118055555555546,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 881.3333333333334,
+      "completions/mean_length": 185.73351542154947,
+      "completions/mean_terminated_length": 157.97284952799478,
+      "completions/min_length": 8.333333333333334,
+      "completions/min_terminated_length": 8.333333333333334,
+      "epoch": 1.725897920604915,
+      "grad_norm": 0.07077532261610031,
+      "kl": 0.07100423177083333,
+      "learning_rate": 1e-06,
+      "loss": 0.0255,
+      "num_tokens": 31578315.0,
+      "reward": 0.6285200913747152,
+      "reward_std": 0.2933768729368846,
+      "rewards/get_embedding_sim/mean": 0.5260895093282064,
+      "rewards/get_embedding_sim/std": 0.10419273873170216,
+      "rewards/reward_num_unique_chars/mean": 0.10243055472771327,
+      "rewards/reward_num_unique_chars/std": 0.30191460251808167,
+      "step": 114
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.035590277777777755,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 998.3333333333334,
+      "completions/mean_length": 197.5555623372396,
+      "completions/mean_terminated_length": 167.07290649414062,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.7712665406427222,
+      "grad_norm": 0.07132314145565033,
+      "kl": 0.07155863444010417,
+      "learning_rate": 1e-06,
+      "loss": 0.0289,
+      "num_tokens": 32370091.0,
+      "reward": 0.6605067054430643,
+      "reward_std": 0.3198150396347046,
+      "rewards/get_embedding_sim/mean": 0.5276941855748495,
+      "rewards/get_embedding_sim/std": 0.09764280170202255,
+      "rewards/reward_num_unique_chars/mean": 0.13281250248352686,
+      "rewards/reward_num_unique_chars/std": 0.3356940845648448,
+      "step": 117
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03819444444444442,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 925.3333333333334,
+      "completions/mean_length": 205.5260467529297,
+      "completions/mean_terminated_length": 172.99127197265625,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.8166351606805293,
+      "grad_norm": 0.07695771753787994,
+      "kl": 0.079559326171875,
+      "learning_rate": 1e-06,
+      "loss": 0.0285,
+      "num_tokens": 33183529.0,
+      "reward": 0.6505021651585897,
+      "reward_std": 0.28806476791699726,
+      "rewards/get_embedding_sim/mean": 0.5255021254221598,
+      "rewards/get_embedding_sim/std": 0.10448584208885829,
+      "rewards/reward_num_unique_chars/mean": 0.12499999751647313,
+      "rewards/reward_num_unique_chars/std": 0.3297826250394185,
+      "step": 120
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03125,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 956.0,
+      "completions/mean_length": 187.78907267252603,
+      "completions/mean_terminated_length": 160.96214803059897,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 1.8620037807183365,
+      "grad_norm": 0.10003960132598877,
+      "kl": 0.10397847493489583,
+      "learning_rate": 1e-06,
+      "loss": 0.0332,
+      "num_tokens": 33979270.0,
+      "reward": 0.702047864596049,
+      "reward_std": 0.2998199959595998,
+      "rewards/get_embedding_sim/mean": 0.5353811780611674,
+      "rewards/get_embedding_sim/std": 0.1009945347905159,
+      "rewards/reward_num_unique_chars/mean": 0.16666666915019354,
+      "rewards/reward_num_unique_chars/std": 0.3635033369064331,
+      "step": 123
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.0390625,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 961.0,
+      "completions/mean_length": 201.6701456705729,
+      "completions/mean_terminated_length": 168.43896484375,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 1.9073724007561437,
+      "grad_norm": 0.1418294459581375,
+      "kl": 0.08981831868489583,
+      "learning_rate": 1e-06,
+      "loss": 0.0311,
+      "num_tokens": 34790042.0,
+      "reward": 0.6395866274833679,
+      "reward_std": 0.278631071249644,
+      "rewards/get_embedding_sim/mean": 0.5371560255686442,
+      "rewards/get_embedding_sim/std": 0.10253078490495682,
+      "rewards/reward_num_unique_chars/mean": 0.10243055721124013,
+      "rewards/reward_num_unique_chars/std": 0.3033109207948049,
+      "step": 126
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.020833333333333297,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 933.6666666666666,
+      "completions/mean_length": 165.38021341959634,
+      "completions/mean_terminated_length": 147.11800384521484,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 1.9527410207939508,
+      "grad_norm": 0.085059255361557,
+      "kl": 0.09361775716145833,
+      "learning_rate": 1e-06,
+      "loss": 0.0241,
+      "num_tokens": 35555504.0,
+      "reward": 0.7251607775688171,
+      "reward_std": 0.3248043159643809,
+      "rewards/get_embedding_sim/mean": 0.5359246134757996,
+      "rewards/get_embedding_sim/std": 0.10831368962923686,
+      "rewards/reward_num_unique_chars/mean": 0.18923610945542654,
+      "rewards/reward_num_unique_chars/std": 0.38598161935806274,
+      "step": 129
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.051843869731800774,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 949.6666666666666,
+      "completions/mean_length": 218.2733408610026,
+      "completions/mean_terminated_length": 174.35783894856772,
+      "completions/min_length": 5.333333333333333,
+      "completions/min_terminated_length": 5.333333333333333,
+      "epoch": 1.998109640831758,
+      "grad_norm": 0.15158401429653168,
+      "kl": 0.09020487467447917,
+      "learning_rate": 1e-06,
+      "loss": 0.0421,
+      "num_tokens": 36367765.0,
+      "reward": 0.6982676188151041,
+      "reward_std": 0.3466052810351054,
+      "rewards/get_embedding_sim/mean": 0.5420175790786743,
+      "rewards/get_embedding_sim/std": 0.09902476519346237,
+      "rewards/reward_num_unique_chars/mean": 0.15625,
+      "rewards/reward_num_unique_chars/std": 0.3612334032853444,
+      "step": 132
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03472222222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 948.0,
+      "completions/mean_length": 176.51909891764322,
+      "completions/mean_terminated_length": 146.023562113444,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.045368620037807,
+      "grad_norm": 0.08893448859453201,
+      "kl": 0.096466064453125,
+      "learning_rate": 1e-06,
+      "loss": 0.0402,
+      "num_tokens": 37147067.0,
+      "reward": 0.7173450986544291,
+      "reward_std": 0.35685937603314716,
+      "rewards/get_embedding_sim/mean": 0.5341853896776835,
+      "rewards/get_embedding_sim/std": 0.1000617394844691,
+      "rewards/reward_num_unique_chars/mean": 0.1831597238779068,
+      "rewards/reward_num_unique_chars/std": 0.3842338224252065,
+      "step": 135
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.041666666666666664,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 879.0,
+      "completions/mean_length": 208.60938008626303,
+      "completions/mean_terminated_length": 173.26571655273438,
+      "completions/min_length": 7.0,
+      "completions/min_terminated_length": 7.0,
+      "epoch": 2.0907372400756143,
+      "grad_norm": 0.12921324372291565,
+      "kl": 0.09910074869791667,
+      "learning_rate": 1e-06,
+      "loss": 0.0475,
+      "num_tokens": 37974473.0,
+      "reward": 0.672684927781423,
+      "reward_std": 0.34854390223821,
+      "rewards/get_embedding_sim/mean": 0.5364001393318176,
+      "rewards/get_embedding_sim/std": 0.10567483057578404,
+      "rewards/reward_num_unique_chars/mean": 0.13628472139437994,
+      "rewards/reward_num_unique_chars/std": 0.34236905972162884,
+      "step": 138
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.021701388888888878,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 956.0,
+      "completions/mean_length": 167.10590616861978,
+      "completions/mean_terminated_length": 148.0250244140625,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.1361058601134215,
+      "grad_norm": 0.12040314823389053,
+      "kl": 0.24815877278645834,
+      "learning_rate": 1e-06,
+      "loss": 0.0403,
+      "num_tokens": 38741491.0,
+      "reward": 0.6958853205045065,
+      "reward_std": 0.3416078786055247,
+      "rewards/get_embedding_sim/mean": 0.5396353205045065,
+      "rewards/get_embedding_sim/std": 0.11144034812847774,
+      "rewards/reward_num_unique_chars/mean": 0.15625000248352686,
+      "rewards/reward_num_unique_chars/std": 0.3600207368532817,
+      "step": 141
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.032986111111111084,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 909.0,
+      "completions/mean_length": 182.28211975097656,
+      "completions/mean_terminated_length": 153.65155029296875,
+      "completions/min_length": 6.0,
+      "completions/min_terminated_length": 6.0,
+      "epoch": 2.1814744801512287,
+      "grad_norm": 0.08720903098583221,
+      "kl": 0.11935933430989583,
+      "learning_rate": 1e-06,
+      "loss": 0.042,
+      "num_tokens": 39531464.0,
+      "reward": 0.7134884198506674,
+      "reward_std": 0.36159368356068927,
+      "rewards/get_embedding_sim/mean": 0.5424814422925314,
+      "rewards/get_embedding_sim/std": 0.11029936373233795,
+      "rewards/reward_num_unique_chars/mean": 0.17100694278875986,
+      "rewards/reward_num_unique_chars/std": 0.37471526861190796,
+      "step": 144
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.03125,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 917.0,
+      "completions/mean_length": 171.84375508626303,
+      "completions/mean_terminated_length": 144.29528299967447,
+      "completions/min_length": 8.0,
+      "completions/min_terminated_length": 8.0,
+      "epoch": 2.226843100189036,
+      "grad_norm": 0.08887135237455368,
+      "kl": 0.118194580078125,
+      "learning_rate": 1e-06,
+      "loss": 0.0381,
+      "num_tokens": 40311428.0,
+      "reward": 0.6788019339243571,
+      "reward_std": 0.33359630902608234,
+      "rewards/get_embedding_sim/mean": 0.5451213518778483,
+      "rewards/get_embedding_sim/std": 0.10193872700134914,
+      "rewards/reward_num_unique_chars/mean": 0.133680559694767,
+      "rewards/reward_num_unique_chars/std": 0.34027015169461566,
+      "step": 147
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.01909722222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 922.6666666666666,
+      "completions/mean_length": 165.07205200195312,
+      "completions/mean_terminated_length": 148.45321146647134,
+      "completions/min_length": 6.0,
+      "completions/min_terminated_length": 6.0,
+      "epoch": 2.272211720226843,
+      "grad_norm": 0.08689926564693451,
+      "kl": 0.149566650390625,
+      "learning_rate": 1e-06,
+      "loss": 0.0324,
+      "num_tokens": 41091415.0,
+      "reward": 0.6884604295094808,
+      "reward_std": 0.3447088996569316,
+      "rewards/get_embedding_sim/mean": 0.5495714743932089,
+      "rewards/get_embedding_sim/std": 0.10353380193312962,
+      "rewards/reward_num_unique_chars/mean": 0.13888888557751974,
+      "rewards/reward_num_unique_chars/std": 0.3459552029768626,
+      "step": 150
+    },
+    {
+      "epoch": 2.31758034026465,
+      "grad_norm": 0.10476606339216232,
+      "learning_rate": 1e-06,
+      "loss": 0.0513,
+      "step": 153
+    },
+    {
+      "epoch": 2.31758034026465,
+      "eval_clip_ratio/high_max": 0.0,
+      "eval_clip_ratio/high_mean": 0.0,
+      "eval_clip_ratio/low_mean": 0.0,
+      "eval_clip_ratio/low_min": 0.0,
+      "eval_clip_ratio/region_mean": 0.0,
+      "eval_completions/clipped_ratio": 0.04687500000000001,
+      "eval_completions/max_length": 856.4464285714286,
+      "eval_completions/max_terminated_length": 614.625,
+      "eval_completions/mean_length": 148.9296919277736,
+      "eval_completions/mean_terminated_length": 106.54870585032872,
+      "eval_completions/min_length": 12.107142857142858,
+      "eval_completions/min_terminated_length": 12.107142857142858,
+      "eval_kl": 0.15039280482700892,
+      "eval_loss": 0.05131923779845238,
+      "eval_num_tokens": 41858572.0,
+      "eval_reward": 0.7319182710988181,
+      "eval_reward_std": 0.39004063113991705,
+      "eval_rewards/get_embedding_sim/mean": 0.5399539640971592,
+      "eval_rewards/get_embedding_sim/std": 0.09657471527212433,
+      "eval_rewards/reward_num_unique_chars/mean": 0.19196428627973156,
+      "eval_rewards/reward_num_unique_chars/std": 0.34904111203338417,
+      "eval_runtime": 1578.4274,
+      "eval_samples_per_second": 0.035,
+      "eval_steps_per_second": 0.001,
+      "step": 153
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.022135416666666668,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 918.8333333333334,
+      "completions/mean_length": 154.00824991861978,
+      "completions/mean_terminated_length": 134.3066151936849,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.3629489603024574,
+      "grad_norm": 0.17861098051071167,
+      "kl": 0.1991424560546875,
+      "learning_rate": 1e-06,
+      "loss": 0.0446,
+      "num_tokens": 42603290.0,
+      "reward": 0.7928757965564728,
+      "reward_std": 0.39940689504146576,
+      "rewards/get_embedding_sim/mean": 0.5420077045758566,
+      "rewards/get_embedding_sim/std": 0.10503626987338066,
+      "rewards/reward_num_unique_chars/mean": 0.2508680547277133,
+      "rewards/reward_num_unique_chars/std": 0.43154530723889667,
+      "step": 156
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.02083333333333337,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 943.6666666666666,
+      "completions/mean_length": 149.3498331705729,
+      "completions/mean_terminated_length": 130.67583719889322,
+      "completions/min_length": 6.0,
+      "completions/min_terminated_length": 6.0,
+      "epoch": 2.4083175803402646,
+      "grad_norm": 0.0997217446565628,
+      "kl": 0.176727294921875,
+      "learning_rate": 1e-06,
+      "loss": 0.0479,
+      "num_tokens": 43362573.0,
+      "reward": 0.7896133859952291,
+      "reward_std": 0.37762073675791424,
+      "rewards/get_embedding_sim/mean": 0.5604466795921326,
+      "rewards/get_embedding_sim/std": 0.10085596889257431,
+      "rewards/reward_num_unique_chars/mean": 0.22916666666666666,
+      "rewards/reward_num_unique_chars/std": 0.417032649119695,
+      "step": 159
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.01996527777777779,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 966.0,
+      "completions/mean_length": 133.52865091959634,
+      "completions/mean_terminated_length": 115.48833719889323,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.4536862003780717,
+      "grad_norm": 0.08974426239728928,
+      "kl": 0.13869730631510416,
+      "learning_rate": 1e-06,
+      "loss": 0.0442,
+      "num_tokens": 44099022.0,
+      "reward": 0.8535909652709961,
+      "reward_std": 0.42868249615033466,
+      "rewards/get_embedding_sim/mean": 0.5332783659299215,
+      "rewards/get_embedding_sim/std": 0.10329846044381459,
+      "rewards/reward_num_unique_chars/mean": 0.3203125,
+      "rewards/reward_num_unique_chars/std": 0.45679094394048053,
+      "step": 162
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.026041666666666668,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 908.0,
+      "completions/mean_length": 141.3715337117513,
+      "completions/mean_terminated_length": 117.76270294189453,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.499054820415879,
+      "grad_norm": 0.10181669145822525,
+      "kl": 0.23414103190104166,
+      "learning_rate": 1e-06,
+      "loss": 0.0402,
+      "num_tokens": 44851706.0,
+      "reward": 0.8000141382217407,
+      "reward_std": 0.3941415250301361,
+      "rewards/get_embedding_sim/mean": 0.5456738670667013,
+      "rewards/get_embedding_sim/std": 0.10725356390078862,
+      "rewards/reward_num_unique_chars/mean": 0.2543402810891469,
+      "rewards/reward_num_unique_chars/std": 0.43460813164711,
+      "step": 165
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.018229166666666703,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 882.6666666666666,
+      "completions/mean_length": 141.09028116861978,
+      "completions/mean_terminated_length": 124.72643280029297,
+      "completions/min_length": 6.666666666666667,
+      "completions/min_terminated_length": 6.666666666666667,
+      "epoch": 2.544423440453686,
+      "grad_norm": 0.08525840193033218,
+      "kl": 0.20921834309895834,
+      "learning_rate": 1e-06,
+      "loss": 0.0455,
+      "num_tokens": 45604066.0,
+      "reward": 0.7868956923484802,
+      "reward_std": 0.40804105003674823,
+      "rewards/get_embedding_sim/mean": 0.5299512147903442,
+      "rewards/get_embedding_sim/std": 0.10723193486531575,
+      "rewards/reward_num_unique_chars/mean": 0.2569444378217061,
+      "rewards/reward_num_unique_chars/std": 0.42565350731213886,
+      "step": 168
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.025173611111111122,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 923.3333333333334,
+      "completions/mean_length": 138.74913533528647,
+      "completions/mean_terminated_length": 116.1558354695638,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.5897920604914932,
+      "grad_norm": 0.6628166437149048,
+      "kl": 0.317626953125,
+      "learning_rate": 1e-06,
+      "loss": 0.0491,
+      "num_tokens": 46345857.0,
+      "reward": 0.8313470085461935,
+      "reward_std": 0.4129582444826762,
+      "rewards/get_embedding_sim/mean": 0.5509650309880575,
+      "rewards/get_embedding_sim/std": 0.09154053280750911,
+      "rewards/reward_num_unique_chars/mean": 0.2803819427887599,
+      "rewards/reward_num_unique_chars/std": 0.4417712489763896,
+      "step": 171
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.01128472222222221,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 772.6666666666666,
+      "completions/mean_length": 111.97396087646484,
+      "completions/mean_terminated_length": 101.59329223632812,
+      "completions/min_length": 6.666666666666667,
+      "completions/min_terminated_length": 6.666666666666667,
+      "epoch": 2.6351606805293004,
+      "grad_norm": 0.09945366531610489,
+      "kl": 0.262237548828125,
+      "learning_rate": 1e-06,
+      "loss": 0.038,
+      "num_tokens": 47054931.0,
+      "reward": 0.8715664744377136,
+      "reward_std": 0.4346109131971995,
+      "rewards/get_embedding_sim/mean": 0.549517830212911,
+      "rewards/get_embedding_sim/std": 0.11452717334032059,
+      "rewards/reward_num_unique_chars/mean": 0.3220486094554265,
+      "rewards/reward_num_unique_chars/std": 0.4634987811247508,
+      "step": 174
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.033854166666666685,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 918.0,
+      "completions/mean_length": 155.64192962646484,
+      "completions/mean_terminated_length": 125.11021041870117,
+      "completions/min_length": 6.0,
+      "completions/min_terminated_length": 6.0,
+      "epoch": 2.6805293005671076,
+      "grad_norm": 0.09938167780637741,
+      "kl": 0.2503814697265625,
+      "learning_rate": 1e-06,
+      "loss": 0.054,
+      "num_tokens": 47795091.0,
+      "reward": 0.7676738500595093,
+      "reward_std": 0.39470958709716797,
+      "rewards/get_embedding_sim/mean": 0.5736633539199829,
+      "rewards/get_embedding_sim/std": 0.09976038336753845,
+      "rewards/reward_num_unique_chars/mean": 0.1940104141831398,
+      "rewards/reward_num_unique_chars/std": 0.3958670049905777,
+      "step": 177
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.017361111111111122,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 878.6666666666666,
+      "completions/mean_length": 116.89757283528645,
+      "completions/mean_terminated_length": 100.84752400716145,
+      "completions/min_length": 6.333333333333333,
+      "completions/min_terminated_length": 6.333333333333333,
+      "epoch": 2.7258979206049148,
+      "grad_norm": 0.1425255984067917,
+      "kl": 0.2775370279947917,
+      "learning_rate": 1e-06,
+      "loss": 0.0481,
+      "num_tokens": 48511037.0,
+      "reward": 0.863362193107605,
+      "reward_std": 0.4587005575497945,
+      "rewards/get_embedding_sim/mean": 0.5473899245262146,
+      "rewards/get_embedding_sim/std": 0.09922760476668675,
+      "rewards/reward_num_unique_chars/mean": 0.3159722288449605,
+      "rewards/reward_num_unique_chars/std": 0.4650394419829051,
+      "step": 180
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.014756944444444456,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 891.3333333333334,
+      "completions/mean_length": 102.27864837646484,
+      "completions/mean_terminated_length": 88.46848042805989,
+      "completions/min_length": 7.0,
+      "completions/min_terminated_length": 7.0,
+      "epoch": 2.7712665406427224,
+      "grad_norm": 0.09843874722719193,
+      "kl": 0.34393310546875,
+      "learning_rate": 1e-06,
+      "loss": 0.0436,
+      "num_tokens": 49204478.0,
+      "reward": 0.8661341269810995,
+      "reward_std": 0.46143727501233417,
+      "rewards/get_embedding_sim/mean": 0.5727312763532003,
+      "rewards/get_embedding_sim/std": 0.11564485480388005,
+      "rewards/reward_num_unique_chars/mean": 0.2934027711550395,
+      "rewards/reward_num_unique_chars/std": 0.4531017243862152,
+      "step": 183
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.01996527777777779,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 754.6666666666666,
+      "completions/mean_length": 122.11806233723958,
+      "completions/mean_terminated_length": 103.76323954264323,
+      "completions/min_length": 5.666666666666667,
+      "completions/min_terminated_length": 5.666666666666667,
+      "epoch": 2.816635160680529,
+      "grad_norm": 0.15066391229629517,
+      "kl": 0.3179728190104167,
+      "learning_rate": 1e-06,
+      "loss": 0.0582,
+      "num_tokens": 49927110.0,
+      "reward": 0.892190178235372,
+      "reward_std": 0.4511215090751648,
+      "rewards/get_embedding_sim/mean": 0.5527804295221964,
+      "rewards/get_embedding_sim/std": 0.1083058441678683,
+      "rewards/reward_num_unique_chars/mean": 0.3394097238779068,
+      "rewards/reward_num_unique_chars/std": 0.46636247634887695,
+      "step": 186
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.022569444444444458,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 940.3333333333334,
+      "completions/mean_length": 126.84115091959636,
+      "completions/mean_terminated_length": 106.15006764729817,
+      "completions/min_length": 7.333333333333333,
+      "completions/min_terminated_length": 7.333333333333333,
+      "epoch": 2.8620037807183367,
+      "grad_norm": 2.990046739578247,
+      "kl": 0.47100830078125,
+      "learning_rate": 1e-06,
+      "loss": 0.0551,
+      "num_tokens": 50663055.0,
+      "reward": 0.8623983860015869,
+      "reward_std": 0.4631191889444987,
+      "rewards/get_embedding_sim/mean": 0.5533705353736877,
+      "rewards/get_embedding_sim/std": 0.11140244205792744,
+      "rewards/reward_num_unique_chars/mean": 0.3090277810891469,
+      "rewards/reward_num_unique_chars/std": 0.4615551829338074,
+      "step": 189
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.01649305555555558,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 891.0,
+      "completions/mean_length": 111.4730936686198,
+      "completions/mean_terminated_length": 96.25564575195312,
+      "completions/min_length": 5.333333333333333,
+      "completions/min_terminated_length": 5.333333333333333,
+      "epoch": 2.9073724007561434,
+      "grad_norm": 0.11676046997308731,
+      "kl": 0.4471232096354167,
+      "learning_rate": 1e-06,
+      "loss": 0.0481,
+      "num_tokens": 51373952.0,
+      "reward": 0.9203431606292725,
+      "reward_std": 0.47053369879722595,
+      "rewards/get_embedding_sim/mean": 0.5444750587145487,
+      "rewards/get_embedding_sim/std": 0.1073705404996872,
+      "rewards/reward_num_unique_chars/mean": 0.3758680522441864,
+      "rewards/reward_num_unique_chars/std": 0.4811862111091614,
+      "step": 192
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.016493055555555542,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 776.3333333333334,
+      "completions/mean_length": 113.12673950195312,
+      "completions/mean_terminated_length": 97.83474731445312,
+      "completions/min_length": 8.666666666666666,
+      "completions/min_terminated_length": 8.666666666666666,
+      "epoch": 2.952741020793951,
+      "grad_norm": 0.09854816645383835,
+      "kl": 0.285675048828125,
+      "learning_rate": 1e-06,
+      "loss": 0.0465,
+      "num_tokens": 52094098.0,
+      "reward": 0.888769249121348,
+      "reward_std": 0.46734312176704407,
+      "rewards/get_embedding_sim/mean": 0.5519636472066244,
+      "rewards/get_embedding_sim/std": 0.11934416989485423,
+      "rewards/reward_num_unique_chars/mean": 0.3368055522441864,
+      "rewards/reward_num_unique_chars/std": 0.4713793396949768,
+      "step": 195
+    },
+    {
+      "clip_ratio/high_max": 0.0,
+      "clip_ratio/high_mean": 0.0,
+      "clip_ratio/low_mean": 0.0,
+      "clip_ratio/low_min": 0.0,
+      "clip_ratio/region_mean": 0.0,
+      "completions/clipped_ratio": 0.013888888888888876,
+      "completions/max_length": 1024.0,
+      "completions/max_terminated_length": 927.3333333333334,
+      "completions/mean_length": 112.50087229410808,
+      "completions/mean_terminated_length": 99.5953369140625,
+      "completions/min_length": 7.0,
+      "completions/min_terminated_length": 7.0,
+      "epoch": 2.998109640831758,
+      "grad_norm": 0.08575434237718582,
+      "kl": 0.42242431640625,
+      "learning_rate": 1e-06,
+      "loss": 0.0494,
+      "num_tokens": 52800707.0,
+      "reward": 0.9310129086176554,
+      "reward_std": 0.4663335382938385,
+      "rewards/get_embedding_sim/mean": 0.5603530804316202,
+      "rewards/get_embedding_sim/std": 0.10822075108687083,
+      "rewards/reward_num_unique_chars/mean": 0.3706597288449605,
+      "rewards/reward_num_unique_chars/std": 0.4820249378681183,
+      "step": 198
+    }
+  ],
+  "logging_steps": 3,
+  "max_steps": 198,
+  "num_input_tokens_seen": 52800707,
+  "num_train_epochs": 3,
+  "save_steps": 25,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e5392c215500566e7a8d9555c9b76802dc16443b9b8ee6f2f9c208e70fe2fbc
+size 8056

zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/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 gc
+import json
+import numpy as np
+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, weights_only=False)
+
+        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 tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # 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}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] 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")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+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 = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * 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
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[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}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        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 to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=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
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    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.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    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")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+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,
+                                                          lazy_mode=True)
+
+    # 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")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_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 by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # 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)