Upload folder using huggingface_hub

added_tokens.json

@@ -0,0 +1,6 @@
+{
+  "<pad>": 151646,
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "/apdcephfs/share_300000800/user/ruotianma/self-correction/saved_models/pre_sft_models/Qwen2-7B-Instruct_sft_qwen2_7B_0123_new_veri_4.5k_20250123_025733/checkpoint-264",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 128245,
+  "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,
+  "pad_token_id": 151646,
+  "rms_norm_eps": 1e-06,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.3",
+  "use_cache": false,
+  "use_sliding_window": false,
+  "vocab_size": 151647
+}

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": 1.05,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.39.3"
+}

latest

@@ -0,0 +1 @@
+global_step152

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ea3d8de4c328cf4e25ef100b43b1c87c76ed03183df80d952700553dff15678
+size 4874671720

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de9fbede83895e5cb720b56849d0911b6d40e03373387aabd17a81ecacdf051c
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0522ff876e2a55110c0430dcaa256b2443a4b9588c80b808109f82918fe04f3c
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:010dc5343dc0cb94401382153575231c4d4a592a2c716e394de8c9241976dadf
+size 1087005824

model.safetensors.index.json

@@ -0,0 +1,346 @@
+{
+  "metadata": {
+    "total_size": 15225254912
+  },
+  "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 sha256:460947adbbb267a1b4cf3d93aa5766a289cdc8f60876efe3c1b4f177d7f321dd
+size 15984

rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5993976e762f12e155b544790780086d8e52904534e8576ede356e2c6642b070
+size 15984

rng_state_10.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f4212887973431f08ed375f9bfa366b7d01a965f573bfb82ee05394b7156a6a
+size 15997

rng_state_11.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d5a26310439da234901a58091c2ddf51dd30f49b113858d6e263b434ea4cc24
+size 15997

rng_state_12.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c4c69d69a8131c7f7a6cb00d3d5d90549176cbc87f2381e5252694b40acfe21
+size 15997

rng_state_13.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e69838ef5f3823a2404592bccdfe5a8c5be816f1011dbf5ae510d1169e27ea9e
+size 15997

rng_state_14.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:266cfbe571a85dd7b8e794be43c738dbb694e3b541d62009450a7ebcfe35b843
+size 15997

rng_state_15.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74c8ec836fd02311db69aed820b895adca7be05a4da774cdbd6b772bc17e303f
+size 15997

rng_state_16.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5bba6034e2a5da3d7ac0b923f1be61d7ae11376a9d274a4fb922488e33633c25
+size 15997

rng_state_17.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96869f2101e2e2f6bafec3d64b5f73b2e55dd3262bfb3b4aaefe0469dd4e9f54
+size 15997

rng_state_18.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c1b20cf4a64e4e4240abb5e9e64ac86bf77aeb6f66f7be622542a442c46a949
+size 15997

rng_state_19.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ea6d99ffe421740269c5ecaa4e3ce2e610c29f27a394eb9c2daea14030c3186
+size 15997

rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c99c0a6a60b2b7235bf06e5ce692d74a7c029aaa273e162330933aa22cc911ff
+size 15984

rng_state_20.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:800bb5171e269bb609bdec588de21ccb58c1a7b76dcb82eb96e0ec74681092ec
+size 15997

rng_state_21.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6eee96f029f426cc97858b6a4543a66bca0c86e87a89744625cc620d8b10f74
+size 15997

rng_state_22.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3039346607c18a918b5c0ba396b8f1ebae8a93acd773f108401f854de98e807
+size 15997

rng_state_23.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d986b3b5952c69d50cc0980ac599d5967a4b6c7d56f0e7f1c0924a92ae2b3d6
+size 15997

rng_state_24.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b8b4307d97259b833d72ecb078fde3b5c2daf633f483a1e49fd968d128481a6
+size 15997

rng_state_25.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27b1c49cdcbcc1d8bef61abb38e37d5aecb79b91262294faa2819f6dcba492ef
+size 15997

rng_state_26.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b45c1590bf62bb88f41bc2d86b577146994d922e1e4a45510d3f2ac669b5520
+size 15997

rng_state_27.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b63346ed51e57cc1e7ac0b03ce7fcfa13c3f2c0aea493a276456e7a28f12fc5
+size 15997

rng_state_28.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14efa258de61b387512281e5b904c67b5ac8ed1bc2dd401f3a2d7240900fb54f
+size 15997

rng_state_29.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5ba7e1c86808ac3dc3e68c02d00e557b189daf5f54f7e3e07d8b1aa373cd35f
+size 15997

rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8131dfd4c66e2b15906309767b2bf919ad566648beb23f2471e229476bb256a2
+size 15984

rng_state_30.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b37d8d83b2db7759bb912b0af6fe5a8c4fc962ebf7a5f729c5e4ba09ff70f166
+size 15997

rng_state_31.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2127b537f71b64205f9043aa588a62bdc0f8bc18f9d847b2028c2b9537d6a3d2
+size 15997

rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:600ed298ec689e55dd450e81e47b6ffa9cfc46adb7563f8abad7e605f27c52f5
+size 15984

rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23385f3e3154112166020322163ff3961f72957f9c1b0ab5c8b090e31ffda686
+size 15984

rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99e7ea329139919e8cbf230a7725b7bec620164a594aa40cb06960064a777aff
+size 15984

rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99d6ca3d66e35a4cb741af9332f2902dbda9a62a290be5b3c696752380f5eac2
+size 15984

rng_state_8.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0f09b79bb157770680cbe8faf2e1b4d7160b91d9f1ea40b5da526be28958fba
+size 15984

rng_state_9.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b212d8d9a712eeb90e2af9eaa86a7becc5750b16792987677a6afc433034ca9
+size 15984

special_tokens_map.json

@@ -0,0 +1,28 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,69 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "128244": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "128245": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "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": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>"
+  ],
+  "bos_token": "<s>",
+  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "model_max_length": 6000,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "truncation_side": "left",
+  "unk_token": "<unk>"
+}

trainer_state.json

@@ -0,0 +1,1389 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3536425767049586,
+  "eval_steps": 500,
+  "global_step": 152,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "importance_ratio": 1.0001533031463623,
+      "kl_div_avg": 0.0022537275217473507,
+      "learning_rate": 0.0,
+      "loss_func": "stage2",
+      "step": 1,
+      "total_loss": 0.20475485920906067
+    },
+    {
+      "epoch": 0.01,
+      "importance_ratio": 1.0000330209732056,
+      "kl_div_avg": 0.0011816158657893538,
+      "learning_rate": 2.153382790366965e-07,
+      "loss_func": "stage2",
+      "step": 2,
+      "total_loss": 0.17304854094982147
+    },
+    {
+      "epoch": 0.01,
+      "importance_ratio": 0.9998020529747009,
+      "kl_div_avg": 0.0012814232613891363,
+      "learning_rate": 3.4130309724299266e-07,
+      "loss_func": "stage2",
+      "step": 3,
+      "total_loss": 0.39294394850730896
+    },
+    {
+      "epoch": 0.01,
+      "importance_ratio": 1.0002059936523438,
+      "kl_div_avg": 0.001280196476727724,
+      "learning_rate": 4.30676558073393e-07,
+      "loss_func": "stage2",
+      "step": 4,
+      "total_loss": 0.2844714820384979
+    },
+    {
+      "epoch": 0.02,
+      "importance_ratio": 1.0000617504119873,
+      "kl_div_avg": 0.0033944130409508944,
+      "learning_rate": 5e-07,
+      "loss_func": "stage2",
+      "step": 5,
+      "total_loss": 0.38719698786735535
+    },
+    {
+      "epoch": 0.02,
+      "importance_ratio": 1.0002291202545166,
+      "kl_div_avg": 0.0006711427122354507,
+      "learning_rate": 5e-07,
+      "loss_func": "stage2",
+      "step": 6,
+      "total_loss": -0.7354744672775269
+    },
+    {
+      "epoch": 0.02,
+      "importance_ratio": 1.0001697540283203,
+      "kl_div_avg": 0.0009293262264691293,
+      "learning_rate": 4.997668997668998e-07,
+      "loss_func": "stage2",
+      "step": 7,
+      "total_loss": 0.10072920471429825
+    },
+    {
+      "epoch": 0.02,
+      "importance_ratio": 1.0003812313079834,
+      "kl_div_avg": 0.0006424246821552515,
+      "learning_rate": 4.995337995337996e-07,
+      "loss_func": "stage2",
+      "step": 8,
+      "total_loss": -0.7356305122375488
+    },
+    {
+      "epoch": 0.03,
+      "importance_ratio": 1.0002334117889404,
+      "kl_div_avg": 0.001079258043318987,
+      "learning_rate": 4.993006993006993e-07,
+      "loss_func": "stage2",
+      "step": 9,
+      "total_loss": 0.1572389006614685
+    },
+    {
+      "epoch": 0.03,
+      "importance_ratio": 0.9999626874923706,
+      "kl_div_avg": 0.0012779454700648785,
+      "learning_rate": 4.990675990675991e-07,
+      "loss_func": "stage2",
+      "step": 10,
+      "total_loss": -0.2823958396911621
+    },
+    {
+      "epoch": 0.03,
+      "importance_ratio": 0.9999264478683472,
+      "kl_div_avg": 0.0010424605570733547,
+      "learning_rate": 4.988344988344988e-07,
+      "loss_func": "stage2",
+      "step": 11,
+      "total_loss": -0.4896008372306824
+    },
+    {
+      "epoch": 0.03,
+      "importance_ratio": 0.9999059438705444,
+      "kl_div_avg": 0.0011010458692908287,
+      "learning_rate": 4.986013986013987e-07,
+      "loss_func": "stage2",
+      "step": 12,
+      "total_loss": 0.4140966832637787
+    },
+    {
+      "epoch": 0.04,
+      "importance_ratio": 0.9999853372573853,
+      "kl_div_avg": 0.0014636358246207237,
+      "learning_rate": 4.983682983682983e-07,
+      "loss_func": "stage2",
+      "step": 13,
+      "total_loss": 0.4479982852935791
+    },
+    {
+      "epoch": 0.04,
+      "importance_ratio": 1.0001029968261719,
+      "kl_div_avg": 0.0010808318620547652,
+      "learning_rate": 4.981351981351981e-07,
+      "loss_func": "stage2",
+      "step": 14,
+      "total_loss": -0.8298860192298889
+    },
+    {
+      "epoch": 0.04,
+      "importance_ratio": 1.0000591278076172,
+      "kl_div_avg": 0.0012179139303043485,
+      "learning_rate": 4.979020979020978e-07,
+      "loss_func": "stage2",
+      "step": 15,
+      "total_loss": -0.8044635057449341
+    },
+    {
+      "epoch": 0.04,
+      "importance_ratio": 1.0000298023223877,
+      "kl_div_avg": 0.004863352049142122,
+      "learning_rate": 4.976689976689976e-07,
+      "loss_func": "stage2",
+      "step": 16,
+      "total_loss": 0.24484601616859436
+    },
+    {
+      "epoch": 0.05,
+      "importance_ratio": 1.0002245903015137,
+      "kl_div_avg": 0.0017589405179023743,
+      "learning_rate": 4.974358974358974e-07,
+      "loss_func": "stage2",
+      "step": 17,
+      "total_loss": -0.0013702064752578735
+    },
+    {
+      "epoch": 0.05,
+      "importance_ratio": 0.9999561309814453,
+      "kl_div_avg": 0.0018663634546101093,
+      "learning_rate": 4.972027972027972e-07,
+      "loss_func": "stage2",
+      "step": 18,
+      "total_loss": -0.10264579951763153
+    },
+    {
+      "epoch": 0.05,
+      "importance_ratio": 0.9998416900634766,
+      "kl_div_avg": 0.0018756331410259008,
+      "learning_rate": 4.969696969696969e-07,
+      "loss_func": "stage2",
+      "step": 19,
+      "total_loss": 0.48283857107162476
+    },
+    {
+      "epoch": 0.05,
+      "importance_ratio": 0.999815821647644,
+      "kl_div_avg": 0.0019035658333450556,
+      "learning_rate": 4.967365967365967e-07,
+      "loss_func": "stage2",
+      "step": 20,
+      "total_loss": -0.1848379671573639
+    },
+    {
+      "epoch": 0.06,
+      "importance_ratio": 0.999942421913147,
+      "kl_div_avg": 0.009937961585819721,
+      "learning_rate": 4.965034965034965e-07,
+      "loss_func": "stage2",
+      "step": 21,
+      "total_loss": 0.14985397458076477
+    },
+    {
+      "epoch": 0.06,
+      "importance_ratio": 0.9997897148132324,
+      "kl_div_avg": 0.00241913297213614,
+      "learning_rate": 4.962703962703962e-07,
+      "loss_func": "stage2",
+      "step": 22,
+      "total_loss": -0.29083502292633057
+    },
+    {
+      "epoch": 0.06,
+      "importance_ratio": 0.9998656511306763,
+      "kl_div_avg": 0.00263982149772346,
+      "learning_rate": 4.96037296037296e-07,
+      "loss_func": "stage2",
+      "step": 23,
+      "total_loss": -0.02688920497894287
+    },
+    {
+      "epoch": 0.06,
+      "importance_ratio": 1.0000394582748413,
+      "kl_div_avg": 0.0032202559523284435,
+      "learning_rate": 4.958041958041958e-07,
+      "loss_func": "stage2",
+      "step": 24,
+      "total_loss": 0.5130484104156494
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 1.0000278949737549,
+      "kl_div_avg": 0.00243174796923995,
+      "learning_rate": 4.955710955710956e-07,
+      "loss_func": "stage2",
+      "step": 25,
+      "total_loss": 0.309948205947876
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9999406337738037,
+      "kl_div_avg": 0.003059545997530222,
+      "learning_rate": 4.953379953379953e-07,
+      "loss_func": "stage2",
+      "step": 26,
+      "total_loss": 0.11305176466703415
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9999201893806458,
+      "kl_div_avg": 0.003822761122137308,
+      "learning_rate": 4.951048951048951e-07,
+      "loss_func": "stage2",
+      "step": 27,
+      "total_loss": 0.38959354162216187
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9994360208511353,
+      "kl_div_avg": 0.0033927513286471367,
+      "learning_rate": 4.948717948717949e-07,
+      "loss_func": "stage2",
+      "step": 28,
+      "total_loss": 0.4653158485889435
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9999792575836182,
+      "kl_div_avg": 0.0032504587434232235,
+      "learning_rate": 4.946386946386946e-07,
+      "loss_func": "stage2",
+      "step": 29,
+      "total_loss": 0.4534304141998291
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 1.0000840425491333,
+      "kl_div_avg": 0.002915366552770138,
+      "learning_rate": 4.944055944055944e-07,
+      "loss_func": "stage2",
+      "step": 30,
+      "total_loss": 0.46045419573783875
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9996820688247681,
+      "kl_div_avg": 0.0030386601574718952,
+      "learning_rate": 4.941724941724942e-07,
+      "loss_func": "stage2",
+      "step": 31,
+      "total_loss": -0.8699095845222473
+    },
+    {
+      "epoch": 0.07,
+      "importance_ratio": 0.9997460842132568,
+      "kl_div_avg": 0.0037004691548645496,
+      "learning_rate": 4.93939393939394e-07,
+      "loss_func": "stage2",
+      "step": 32,
+      "total_loss": -0.23050172626972198
+    },
+    {
+      "epoch": 0.08,
+      "importance_ratio": 0.9999768137931824,
+      "kl_div_avg": 0.0046552978456020355,
+      "learning_rate": 4.937062937062936e-07,
+      "loss_func": "stage2",
+      "step": 33,
+      "total_loss": 0.3776797950267792
+    },
+    {
+      "epoch": 0.08,
+      "importance_ratio": 0.9999284148216248,
+      "kl_div_avg": 0.004839582834392786,
+      "learning_rate": 4.934731934731934e-07,
+      "loss_func": "stage2",
+      "step": 34,
+      "total_loss": 0.5804722309112549
+    },
+    {
+      "epoch": 0.08,
+      "importance_ratio": 0.999695897102356,
+      "kl_div_avg": 0.004378842655569315,
+      "learning_rate": 4.932400932400932e-07,
+      "loss_func": "stage2",
+      "step": 35,
+      "total_loss": 0.5690972805023193
+    },
+    {
+      "epoch": 0.08,
+      "importance_ratio": 0.9998815059661865,
+      "kl_div_avg": 0.0047516971826553345,
+      "learning_rate": 4.93006993006993e-07,
+      "loss_func": "stage2",
+      "step": 36,
+      "total_loss": 0.2298603653907776
+    },
+    {
+      "epoch": 0.09,
+      "importance_ratio": 0.9997518062591553,
+      "kl_div_avg": 0.004270514938980341,
+      "learning_rate": 4.927738927738927e-07,
+      "loss_func": "stage2",
+      "step": 37,
+      "total_loss": -0.20016932487487793
+    },
+    {
+      "epoch": 0.09,
+      "importance_ratio": 1.000083088874817,
+      "kl_div_avg": 0.004711843561381102,
+      "learning_rate": 4.925407925407925e-07,
+      "loss_func": "stage2",
+      "step": 38,
+      "total_loss": 0.25743457674980164
+    },
+    {
+      "epoch": 0.09,
+      "importance_ratio": 0.9999402761459351,
+      "kl_div_avg": 0.004922826308757067,
+      "learning_rate": 4.923076923076923e-07,
+      "loss_func": "stage2",
+      "step": 39,
+      "total_loss": -0.15881219506263733
+    },
+    {
+      "epoch": 0.09,
+      "importance_ratio": 0.999858021736145,
+      "kl_div_avg": 0.0039229318499565125,
+      "learning_rate": 4.92074592074592e-07,
+      "loss_func": "stage2",
+      "step": 40,
+      "total_loss": -0.23230578005313873
+    },
+    {
+      "epoch": 0.1,
+      "importance_ratio": 0.999944806098938,
+      "kl_div_avg": 0.00766429677605629,
+      "learning_rate": 4.918414918414918e-07,
+      "loss_func": "stage2",
+      "step": 41,
+      "total_loss": -0.03111131489276886
+    },
+    {
+      "epoch": 0.1,
+      "importance_ratio": 1.000047206878662,
+      "kl_div_avg": 0.005274048075079918,
+      "learning_rate": 4.916083916083916e-07,
+      "loss_func": "stage2",
+      "step": 42,
+      "total_loss": -0.033877044916152954
+    },
+    {
+      "epoch": 0.1,
+      "importance_ratio": 1.0002098083496094,
+      "kl_div_avg": 0.006382378749549389,
+      "learning_rate": 4.913752913752914e-07,
+      "loss_func": "stage2",
+      "step": 43,
+      "total_loss": 0.10409477353096008
+    },
+    {
+      "epoch": 0.1,
+      "importance_ratio": 0.9998437166213989,
+      "kl_div_avg": 0.006663881242275238,
+      "learning_rate": 4.911421911421911e-07,
+      "loss_func": "stage2",
+      "step": 44,
+      "total_loss": -0.9234535694122314
+    },
+    {
+      "epoch": 0.11,
+      "importance_ratio": 0.9999840259552002,
+      "kl_div_avg": 0.004546988755464554,
+      "learning_rate": 4.909090909090909e-07,
+      "loss_func": "stage2",
+      "step": 45,
+      "total_loss": 0.022589027881622314
+    },
+    {
+      "epoch": 0.11,
+      "importance_ratio": 1.0002349615097046,
+      "kl_div_avg": 0.0048853568732738495,
+      "learning_rate": 4.906759906759906e-07,
+      "loss_func": "stage2",
+      "step": 46,
+      "total_loss": 0.7868871688842773
+    },
+    {
+      "epoch": 0.11,
+      "importance_ratio": 1.0003743171691895,
+      "kl_div_avg": 0.005253675393760204,
+      "learning_rate": 4.904428904428905e-07,
+      "loss_func": "stage2",
+      "step": 47,
+      "total_loss": 0.7918493747711182
+    },
+    {
+      "epoch": 0.11,
+      "importance_ratio": 1.0001533031463623,
+      "kl_div_avg": 0.005680109839886427,
+      "learning_rate": 4.902097902097902e-07,
+      "loss_func": "stage2",
+      "step": 48,
+      "total_loss": -0.10262584686279297
+    },
+    {
+      "epoch": 0.12,
+      "importance_ratio": 0.999789834022522,
+      "kl_div_avg": 0.006105936132371426,
+      "learning_rate": 4.8997668997669e-07,
+      "loss_func": "stage2",
+      "step": 49,
+      "total_loss": -0.8303477168083191
+    },
+    {
+      "epoch": 0.12,
+      "importance_ratio": 0.9999826550483704,
+      "kl_div_avg": 0.005722599104046822,
+      "learning_rate": 4.897435897435897e-07,
+      "loss_func": "stage2",
+      "step": 50,
+      "total_loss": -0.8298835754394531
+    },
+    {
+      "epoch": 0.12,
+      "importance_ratio": 1.0003968477249146,
+      "kl_div_avg": 0.005780387669801712,
+      "learning_rate": 4.895104895104895e-07,
+      "loss_func": "stage2",
+      "step": 51,
+      "total_loss": -0.8298872709274292
+    },
+    {
+      "epoch": 0.12,
+      "importance_ratio": 1.001389741897583,
+      "kl_div_avg": 0.00528342742472887,
+      "learning_rate": 4.892773892773893e-07,
+      "loss_func": "stage2",
+      "step": 52,
+      "total_loss": -0.8320926427841187
+    },
+    {
+      "epoch": 0.13,
+      "importance_ratio": 1.0000613927841187,
+      "kl_div_avg": 0.005511891096830368,
+      "learning_rate": 4.890442890442891e-07,
+      "loss_func": "stage2",
+      "step": 53,
+      "total_loss": -0.8287703990936279
+    },
+    {
+      "epoch": 0.13,
+      "importance_ratio": 1.000309705734253,
+      "kl_div_avg": 0.00582331046462059,
+      "learning_rate": 4.888111888111888e-07,
+      "loss_func": "stage2",
+      "step": 54,
+      "total_loss": -0.007962286472320557
+    },
+    {
+      "epoch": 0.13,
+      "importance_ratio": 1.000248670578003,
+      "kl_div_avg": 0.005605565384030342,
+      "learning_rate": 4.885780885780885e-07,
+      "loss_func": "stage2",
+      "step": 55,
+      "total_loss": 0.30858537554740906
+    },
+    {
+      "epoch": 0.13,
+      "importance_ratio": 1.0001593828201294,
+      "kl_div_avg": 0.006694035604596138,
+      "learning_rate": 4.883449883449883e-07,
+      "loss_func": "stage2",
+      "step": 56,
+      "total_loss": 0.2841358482837677
+    },
+    {
+      "epoch": 0.14,
+      "importance_ratio": 0.9998378753662109,
+      "kl_div_avg": 0.0054409438744187355,
+      "learning_rate": 4.88111888111888e-07,
+      "loss_func": "stage2",
+      "step": 57,
+      "total_loss": 0.4940628409385681
+    },
+    {
+      "epoch": 0.14,
+      "importance_ratio": 1.0000338554382324,
+      "kl_div_avg": 0.00452791154384613,
+      "learning_rate": 4.878787878787878e-07,
+      "loss_func": "stage2",
+      "step": 58,
+      "total_loss": 0.25203195214271545
+    },
+    {
+      "epoch": 0.14,
+      "importance_ratio": 1.0000851154327393,
+      "kl_div_avg": 0.008046677336096764,
+      "learning_rate": 4.876456876456876e-07,
+      "loss_func": "stage2",
+      "step": 59,
+      "total_loss": 0.8340111970901489
+    },
+    {
+      "epoch": 0.14,
+      "importance_ratio": 1.0000518560409546,
+      "kl_div_avg": 0.0053115119226276875,
+      "learning_rate": 4.874125874125874e-07,
+      "loss_func": "stage2",
+      "step": 60,
+      "total_loss": -0.4959676265716553
+    },
+    {
+      "epoch": 0.15,
+      "importance_ratio": 1.0000547170639038,
+      "kl_div_avg": 0.009003904648125172,
+      "learning_rate": 4.871794871794871e-07,
+      "loss_func": "stage2",
+      "step": 61,
+      "total_loss": -0.7978946566581726
+    },
+    {
+      "epoch": 0.15,
+      "importance_ratio": 1.000227689743042,
+      "kl_div_avg": 0.007788301911205053,
+      "learning_rate": 4.869463869463869e-07,
+      "loss_func": "stage2",
+      "step": 62,
+      "total_loss": -0.801626443862915
+    },
+    {
+      "epoch": 0.15,
+      "importance_ratio": 1.0001925230026245,
+      "kl_div_avg": 0.009508013725280762,
+      "learning_rate": 4.867132867132867e-07,
+      "loss_func": "stage2",
+      "step": 63,
+      "total_loss": -0.8087908029556274
+    },
+    {
+      "epoch": 0.15,
+      "importance_ratio": 1.0004205703735352,
+      "kl_div_avg": 0.00948132760822773,
+      "learning_rate": 4.864801864801865e-07,
+      "loss_func": "stage2",
+      "step": 64,
+      "total_loss": -0.7977355718612671
+    },
+    {
+      "epoch": 0.16,
+      "importance_ratio": 0.9995088577270508,
+      "kl_div_avg": 0.013071361929178238,
+      "learning_rate": 4.862470862470862e-07,
+      "loss_func": "stage2",
+      "step": 65,
+      "total_loss": -0.7926455736160278
+    },
+    {
+      "epoch": 0.16,
+      "importance_ratio": 1.0000696182250977,
+      "kl_div_avg": 0.008261503651738167,
+      "learning_rate": 4.86013986013986e-07,
+      "loss_func": "stage2",
+      "step": 66,
+      "total_loss": 0.8929388523101807
+    },
+    {
+      "epoch": 0.16,
+      "importance_ratio": 0.9999169111251831,
+      "kl_div_avg": 0.0072638243436813354,
+      "learning_rate": 4.857808857808858e-07,
+      "loss_func": "stage2",
+      "step": 67,
+      "total_loss": 0.7651723623275757
+    },
+    {
+      "epoch": 0.16,
+      "importance_ratio": 1.0006506443023682,
+      "kl_div_avg": 0.014501883648335934,
+      "learning_rate": 4.855477855477855e-07,
+      "loss_func": "stage2",
+      "step": 68,
+      "total_loss": -0.7893998622894287
+    },
+    {
+      "epoch": 0.17,
+      "importance_ratio": 1.0001804828643799,
+      "kl_div_avg": 0.014889956451952457,
+      "learning_rate": 4.853146853146853e-07,
+      "loss_func": "stage2",
+      "step": 69,
+      "total_loss": -0.28294438123703003
+    },
+    {
+      "epoch": 0.17,
+      "importance_ratio": 1.0003015995025635,
+      "kl_div_avg": 0.012897053733468056,
+      "learning_rate": 4.850815850815851e-07,
+      "loss_func": "stage2",
+      "step": 70,
+      "total_loss": -0.29949530959129333
+    },
+    {
+      "epoch": 0.17,
+      "importance_ratio": 0.9999880194664001,
+      "kl_div_avg": 0.011545092798769474,
+      "learning_rate": 4.848484848484849e-07,
+      "loss_func": "stage2",
+      "step": 71,
+      "total_loss": -0.07207685708999634
+    },
+    {
+      "epoch": 0.17,
+      "importance_ratio": 1.0000548362731934,
+      "kl_div_avg": 0.011798446998000145,
+      "learning_rate": 4.846153846153846e-07,
+      "loss_func": "stage2",
+      "step": 72,
+      "total_loss": -0.030968129634857178
+    },
+    {
+      "epoch": 0.18,
+      "importance_ratio": 0.9998947381973267,
+      "kl_div_avg": 0.014826457016170025,
+      "learning_rate": 4.843822843822844e-07,
+      "loss_func": "stage2",
+      "step": 73,
+      "total_loss": -0.04581499099731445
+    },
+    {
+      "epoch": 0.18,
+      "importance_ratio": 0.9999833106994629,
+      "kl_div_avg": 0.01308258343487978,
+      "learning_rate": 4.841491841491842e-07,
+      "loss_func": "stage2",
+      "step": 74,
+      "total_loss": -0.2057284414768219
+    },
+    {
+      "epoch": 0.18,
+      "importance_ratio": 1.000425100326538,
+      "kl_div_avg": 0.014183840714395046,
+      "learning_rate": 4.839160839160839e-07,
+      "loss_func": "stage2",
+      "step": 75,
+      "total_loss": 0.6338366866111755
+    },
+    {
+      "epoch": 0.18,
+      "importance_ratio": 1.0001928806304932,
+      "kl_div_avg": 0.013401240110397339,
+      "learning_rate": 4.836829836829837e-07,
+      "loss_func": "stage2",
+      "step": 76,
+      "total_loss": 0.2874578833580017
+    },
+    {
+      "epoch": 0.19,
+      "importance_ratio": 0.9999673366546631,
+      "kl_div_avg": 0.012778308242559433,
+      "learning_rate": 4.834498834498834e-07,
+      "loss_func": "stage2",
+      "step": 77,
+      "total_loss": 0.8181835412979126
+    },
+    {
+      "epoch": 0.19,
+      "importance_ratio": 0.9999823570251465,
+      "kl_div_avg": 0.014732200652360916,
+      "learning_rate": 4.832167832167832e-07,
+      "loss_func": "stage2",
+      "step": 78,
+      "total_loss": 0.818813681602478
+    },
+    {
+      "epoch": 0.19,
+      "importance_ratio": 1.0003037452697754,
+      "kl_div_avg": 0.014829716645181179,
+      "learning_rate": 4.829836829836829e-07,
+      "loss_func": "stage2",
+      "step": 79,
+      "total_loss": 0.0298653244972229
+    },
+    {
+      "epoch": 0.19,
+      "importance_ratio": 1.0003316402435303,
+      "kl_div_avg": 0.01861473172903061,
+      "learning_rate": 4.827505827505827e-07,
+      "loss_func": "stage2",
+      "step": 80,
+      "total_loss": 0.47803181409835815
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 1.0000956058502197,
+      "kl_div_avg": 0.013490747660398483,
+      "learning_rate": 4.825174825174824e-07,
+      "loss_func": "stage2",
+      "step": 81,
+      "total_loss": 0.43567734956741333
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 1.0001178979873657,
+      "kl_div_avg": 0.014975698664784431,
+      "learning_rate": 4.822843822843823e-07,
+      "loss_func": "stage2",
+      "step": 82,
+      "total_loss": -0.27731338143348694
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 0.9999943971633911,
+      "kl_div_avg": 0.012739075347781181,
+      "learning_rate": 4.82051282051282e-07,
+      "loss_func": "stage2",
+      "step": 83,
+      "total_loss": 0.14717233180999756
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 1.000013828277588,
+      "kl_div_avg": 0.015260843560099602,
+      "learning_rate": 4.818181818181818e-07,
+      "loss_func": "stage2",
+      "step": 84,
+      "total_loss": -0.876125693321228
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 0.9999637603759766,
+      "kl_div_avg": 0.009859403595328331,
+      "learning_rate": 4.815850815850815e-07,
+      "loss_func": "stage2",
+      "step": 85,
+      "total_loss": -0.007574997842311859
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 0.9998883008956909,
+      "kl_div_avg": 0.012243506498634815,
+      "learning_rate": 4.813519813519813e-07,
+      "loss_func": "stage2",
+      "step": 86,
+      "total_loss": -0.8214981555938721
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 0.9997843503952026,
+      "kl_div_avg": 0.009431181475520134,
+      "learning_rate": 4.811188811188811e-07,
+      "loss_func": "stage2",
+      "step": 87,
+      "total_loss": 0.09674309194087982
+    },
+    {
+      "epoch": 0.2,
+      "importance_ratio": 0.9998980760574341,
+      "kl_div_avg": 0.012359343469142914,
+      "learning_rate": 4.808857808857809e-07,
+      "loss_func": "stage2",
+      "step": 88,
+      "total_loss": 0.13037657737731934
+    },
+    {
+      "epoch": 0.21,
+      "importance_ratio": 0.9999622702598572,
+      "kl_div_avg": 0.019689181819558144,
+      "learning_rate": 4.806526806526806e-07,
+      "loss_func": "stage2",
+      "step": 89,
+      "total_loss": -0.053394585847854614
+    },
+    {
+      "epoch": 0.21,
+      "importance_ratio": 1.0000618696212769,
+      "kl_div_avg": 0.013591418042778969,
+      "learning_rate": 4.804195804195804e-07,
+      "loss_func": "stage2",
+      "step": 90,
+      "total_loss": 0.7699258327484131
+    },
+    {
+      "epoch": 0.21,
+      "importance_ratio": 1.000166893005371,
+      "kl_div_avg": 0.016464080661535263,
+      "learning_rate": 4.801864801864802e-07,
+      "loss_func": "stage2",
+      "step": 91,
+      "total_loss": 0.16640010476112366
+    },
+    {
+      "epoch": 0.21,
+      "importance_ratio": 0.9997192621231079,
+      "kl_div_avg": 0.02152731642127037,
+      "learning_rate": 4.799533799533799e-07,
+      "loss_func": "stage2",
+      "step": 92,
+      "total_loss": -0.17338214814662933
+    },
+    {
+      "epoch": 0.22,
+      "importance_ratio": 1.000048279762268,
+      "kl_div_avg": 0.01632719114422798,
+      "learning_rate": 4.797202797202797e-07,
+      "loss_func": "stage2",
+      "step": 93,
+      "total_loss": -0.311355322599411
+    },
+    {
+      "epoch": 0.22,
+      "importance_ratio": 1.000006914138794,
+      "kl_div_avg": 0.02427072264254093,
+      "learning_rate": 4.794871794871795e-07,
+      "loss_func": "stage2",
+      "step": 94,
+      "total_loss": 0.6689386963844299
+    },
+    {
+      "epoch": 0.22,
+      "importance_ratio": 1.0001646280288696,
+      "kl_div_avg": 0.020599162206053734,
+      "learning_rate": 4.792540792540793e-07,
+      "loss_func": "stage2",
+      "step": 95,
+      "total_loss": -0.27341556549072266
+    },
+    {
+      "epoch": 0.22,
+      "importance_ratio": 1.0005509853363037,
+      "kl_div_avg": 0.01824137195944786,
+      "learning_rate": 4.79020979020979e-07,
+      "loss_func": "stage2",
+      "step": 96,
+      "total_loss": -0.7693477869033813
+    },
+    {
+      "epoch": 0.23,
+      "importance_ratio": 1.0000276565551758,
+      "kl_div_avg": 0.019494881853461266,
+      "learning_rate": 4.787878787878788e-07,
+      "loss_func": "stage2",
+      "step": 97,
+      "total_loss": 0.5362880229949951
+    },
+    {
+      "epoch": 0.23,
+      "importance_ratio": 1.0001310110092163,
+      "kl_div_avg": 0.018007826060056686,
+      "learning_rate": 4.785547785547786e-07,
+      "loss_func": "stage2",
+      "step": 98,
+      "total_loss": 0.13775774836540222
+    },
+    {
+      "epoch": 0.23,
+      "importance_ratio": 1.0001639127731323,
+      "kl_div_avg": 0.018012849614024162,
+      "learning_rate": 4.783216783216783e-07,
+      "loss_func": "stage2",
+      "step": 99,
+      "total_loss": 0.43274223804473877
+    },
+    {
+      "epoch": 0.23,
+      "importance_ratio": 1.0001168251037598,
+      "kl_div_avg": 0.016401609405875206,
+      "learning_rate": 4.78088578088578e-07,
+      "loss_func": "stage2",
+      "step": 100,
+      "total_loss": 0.024271167814731598
+    },
+    {
+      "epoch": 0.24,
+      "importance_ratio": 0.9998679161071777,
+      "kl_div_avg": 0.018232179805636406,
+      "learning_rate": 4.778554778554778e-07,
+      "loss_func": "stage2",
+      "step": 101,
+      "total_loss": -0.13439278304576874
+    },
+    {
+      "epoch": 0.24,
+      "importance_ratio": 0.9999688267707825,
+      "kl_div_avg": 0.01988252066075802,
+      "learning_rate": 4.776223776223776e-07,
+      "loss_func": "stage2",
+      "step": 102,
+      "total_loss": -0.5875260233879089
+    },
+    {
+      "epoch": 0.24,
+      "importance_ratio": 1.000270128250122,
+      "kl_div_avg": 0.024290431290864944,
+      "learning_rate": 4.773892773892773e-07,
+      "loss_func": "stage2",
+      "step": 103,
+      "total_loss": 0.43479496240615845
+    },
+    {
+      "epoch": 0.24,
+      "importance_ratio": 1.0000674724578857,
+      "kl_div_avg": 0.017674673348665237,
+      "learning_rate": 4.771561771561771e-07,
+      "loss_func": "stage2",
+      "step": 104,
+      "total_loss": -0.30525317788124084
+    },
+    {
+      "epoch": 0.25,
+      "importance_ratio": 0.9998775720596313,
+      "kl_div_avg": 0.015626681968569756,
+      "learning_rate": 4.769230769230769e-07,
+      "loss_func": "stage2",
+      "step": 105,
+      "total_loss": -0.4420226514339447
+    },
+    {
+      "epoch": 0.25,
+      "importance_ratio": 1.000077247619629,
+      "kl_div_avg": 0.011648900806903839,
+      "learning_rate": 4.7668997668997666e-07,
+      "loss_func": "stage2",
+      "step": 106,
+      "total_loss": -0.7798103094100952
+    },
+    {
+      "epoch": 0.25,
+      "importance_ratio": 0.9998931884765625,
+      "kl_div_avg": 0.02424587681889534,
+      "learning_rate": 4.7645687645687646e-07,
+      "loss_func": "stage2",
+      "step": 107,
+      "total_loss": 0.3463074564933777
+    },
+    {
+      "epoch": 0.25,
+      "importance_ratio": 1.0001239776611328,
+      "kl_div_avg": 0.019622065126895905,
+      "learning_rate": 4.762237762237762e-07,
+      "loss_func": "stage2",
+      "step": 108,
+      "total_loss": -0.7446590662002563
+    },
+    {
+      "epoch": 0.26,
+      "importance_ratio": 1.0000584125518799,
+      "kl_div_avg": 0.02028917521238327,
+      "learning_rate": 4.75990675990676e-07,
+      "loss_func": "stage2",
+      "step": 109,
+      "total_loss": -0.269249826669693
+    },
+    {
+      "epoch": 0.26,
+      "importance_ratio": 0.9998422861099243,
+      "kl_div_avg": 0.022249866276979446,
+      "learning_rate": 4.7575757575757574e-07,
+      "loss_func": "stage2",
+      "step": 110,
+      "total_loss": -0.6871204376220703
+    },
+    {
+      "epoch": 0.26,
+      "importance_ratio": 1.0001263618469238,
+      "kl_div_avg": 0.01973864436149597,
+      "learning_rate": 4.755244755244755e-07,
+      "loss_func": "stage2",
+      "step": 111,
+      "total_loss": -0.6550001502037048
+    },
+    {
+      "epoch": 0.26,
+      "importance_ratio": 0.9996439218521118,
+      "kl_div_avg": 0.02077101171016693,
+      "learning_rate": 4.7529137529137523e-07,
+      "loss_func": "stage2",
+      "step": 112,
+      "total_loss": 0.042905211448669434
+    },
+    {
+      "epoch": 0.27,
+      "importance_ratio": 1.000232219696045,
+      "kl_div_avg": 0.036871857941150665,
+      "learning_rate": 4.750582750582751e-07,
+      "loss_func": "stage2",
+      "step": 113,
+      "total_loss": -0.6672766804695129
+    },
+    {
+      "epoch": 0.27,
+      "importance_ratio": 0.9999071359634399,
+      "kl_div_avg": 0.023731358349323273,
+      "learning_rate": 4.748251748251748e-07,
+      "loss_func": "stage2",
+      "step": 114,
+      "total_loss": 0.22221048176288605
+    },
+    {
+      "epoch": 0.27,
+      "importance_ratio": 0.9998910427093506,
+      "kl_div_avg": 0.02947179228067398,
+      "learning_rate": 4.7459207459207457e-07,
+      "loss_func": "stage2",
+      "step": 115,
+      "total_loss": 0.05969160795211792
+    },
+    {
+      "epoch": 0.27,
+      "importance_ratio": 0.9999151825904846,
+      "kl_div_avg": 0.028714872896671295,
+      "learning_rate": 4.743589743589743e-07,
+      "loss_func": "stage2",
+      "step": 116,
+      "total_loss": 0.023228317499160767
+    },
+    {
+      "epoch": 0.28,
+      "importance_ratio": 0.9999186992645264,
+      "kl_div_avg": 0.022106900811195374,
+      "learning_rate": 4.741258741258741e-07,
+      "loss_func": "stage2",
+      "step": 117,
+      "total_loss": -0.4819675087928772
+    },
+    {
+      "epoch": 0.28,
+      "importance_ratio": 0.9995689392089844,
+      "kl_div_avg": 0.03649330139160156,
+      "learning_rate": 4.7389277389277386e-07,
+      "loss_func": "stage2",
+      "step": 118,
+      "total_loss": 0.3108961582183838
+    },
+    {
+      "epoch": 0.28,
+      "importance_ratio": 0.9996166825294495,
+      "kl_div_avg": 0.03707721084356308,
+      "learning_rate": 4.7365967365967365e-07,
+      "loss_func": "stage2",
+      "step": 119,
+      "total_loss": 0.3411310315132141
+    },
+    {
+      "epoch": 0.28,
+      "importance_ratio": 0.9993818998336792,
+      "kl_div_avg": 0.02660643495619297,
+      "learning_rate": 4.734265734265734e-07,
+      "loss_func": "stage2",
+      "step": 120,
+      "total_loss": -0.1782078891992569
+    },
+    {
+      "epoch": 0.29,
+      "importance_ratio": 1.0000483989715576,
+      "kl_div_avg": 0.020809054374694824,
+      "learning_rate": 4.731934731934732e-07,
+      "loss_func": "stage2",
+      "step": 121,
+      "total_loss": 0.2691134810447693
+    },
+    {
+      "epoch": 0.29,
+      "importance_ratio": 0.9999549388885498,
+      "kl_div_avg": 0.018363434821367264,
+      "learning_rate": 4.7296037296037294e-07,
+      "loss_func": "stage2",
+      "step": 122,
+      "total_loss": -0.11637084186077118
+    },
+    {
+      "epoch": 0.29,
+      "importance_ratio": 1.0000113248825073,
+      "kl_div_avg": 0.01843072474002838,
+      "learning_rate": 4.727272727272727e-07,
+      "loss_func": "stage2",
+      "step": 123,
+      "total_loss": 0.41335129737854004
+    },
+    {
+      "epoch": 0.29,
+      "importance_ratio": 0.9993541836738586,
+      "kl_div_avg": 0.023193594068288803,
+      "learning_rate": 4.724941724941724e-07,
+      "loss_func": "stage2",
+      "step": 124,
+      "total_loss": 0.3763417601585388
+    },
+    {
+      "epoch": 0.3,
+      "importance_ratio": 1.0000863075256348,
+      "kl_div_avg": 0.02424781210720539,
+      "learning_rate": 4.722610722610723e-07,
+      "loss_func": "stage2",
+      "step": 125,
+      "total_loss": 0.8012444972991943
+    },
+    {
+      "epoch": 0.3,
+      "importance_ratio": 0.9997611045837402,
+      "kl_div_avg": 0.019796304404735565,
+      "learning_rate": 4.72027972027972e-07,
+      "loss_func": "stage2",
+      "step": 126,
+      "total_loss": 0.2471800446510315
+    },
+    {
+      "epoch": 0.3,
+      "importance_ratio": 0.9994201064109802,
+      "kl_div_avg": 0.03692193701863289,
+      "learning_rate": 4.7179487179487176e-07,
+      "loss_func": "stage2",
+      "step": 127,
+      "total_loss": 0.6575199961662292
+    },
+    {
+      "epoch": 0.3,
+      "importance_ratio": 0.9993878602981567,
+      "kl_div_avg": 0.024065542966127396,
+      "learning_rate": 4.715617715617715e-07,
+      "loss_func": "stage2",
+      "step": 128,
+      "total_loss": 0.39468204975128174
+    },
+    {
+      "epoch": 0.31,
+      "importance_ratio": 1.0001311302185059,
+      "kl_div_avg": 0.024539019912481308,
+      "learning_rate": 4.713286713286713e-07,
+      "loss_func": "stage2",
+      "step": 129,
+      "total_loss": 0.1832764893770218
+    },
+    {
+      "epoch": 0.31,
+      "importance_ratio": 0.9998582005500793,
+      "kl_div_avg": 0.024634480476379395,
+      "learning_rate": 4.710955710955711e-07,
+      "loss_func": "stage2",
+      "step": 130,
+      "total_loss": 0.1833437830209732
+    },
+    {
+      "epoch": 0.31,
+      "importance_ratio": 0.9997899532318115,
+      "kl_div_avg": 0.024798087775707245,
+      "learning_rate": 4.7086247086247085e-07,
+      "loss_func": "stage2",
+      "step": 131,
+      "total_loss": -0.029144808650016785
+    },
+    {
+      "epoch": 0.31,
+      "importance_ratio": 0.9991902112960815,
+      "kl_div_avg": 0.022448930889368057,
+      "learning_rate": 4.706293706293706e-07,
+      "loss_func": "stage2",
+      "step": 132,
+      "total_loss": 0.5746316909790039
+    },
+    {
+      "epoch": 0.32,
+      "importance_ratio": 0.9999872446060181,
+      "kl_div_avg": 0.030649660155177116,
+      "learning_rate": 4.703962703962704e-07,
+      "loss_func": "stage2",
+      "step": 133,
+      "total_loss": -0.27921533584594727
+    },
+    {
+      "epoch": 0.32,
+      "importance_ratio": 0.9999322891235352,
+      "kl_div_avg": 0.028502434492111206,
+      "learning_rate": 4.7016317016317013e-07,
+      "loss_func": "stage2",
+      "step": 134,
+      "total_loss": -0.18240980803966522
+    },
+    {
+      "epoch": 0.32,
+      "importance_ratio": 0.9991711378097534,
+      "kl_div_avg": 0.034727346152067184,
+      "learning_rate": 4.699300699300699e-07,
+      "loss_func": "stage2",
+      "step": 135,
+      "total_loss": 0.20225152373313904
+    },
+    {
+      "epoch": 0.32,
+      "importance_ratio": 1.000333547592163,
+      "kl_div_avg": 0.02375342883169651,
+      "learning_rate": 4.696969696969697e-07,
+      "loss_func": "stage2",
+      "step": 136,
+      "total_loss": 0.19642743468284607
+    },
+    {
+      "epoch": 0.33,
+      "importance_ratio": 1.0000425577163696,
+      "kl_div_avg": 0.02941157855093479,
+      "learning_rate": 4.6946386946386947e-07,
+      "loss_func": "stage2",
+      "step": 137,
+      "total_loss": -0.2832520604133606
+    },
+    {
+      "epoch": 0.33,
+      "importance_ratio": 0.9998887777328491,
+      "kl_div_avg": 0.0380094014108181,
+      "learning_rate": 4.692307692307692e-07,
+      "loss_func": "stage2",
+      "step": 138,
+      "total_loss": 0.07638365030288696
+    },
+    {
+      "epoch": 0.33,
+      "importance_ratio": 1.0000771284103394,
+      "kl_div_avg": 0.031544946134090424,
+      "learning_rate": 4.6899766899766896e-07,
+      "loss_func": "stage2",
+      "step": 139,
+      "total_loss": -0.29178526997566223
+    },
+    {
+      "epoch": 0.33,
+      "importance_ratio": 0.9999626278877258,
+      "kl_div_avg": 0.03620228171348572,
+      "learning_rate": 4.6876456876456875e-07,
+      "loss_func": "stage2",
+      "step": 140,
+      "total_loss": 0.12343016266822815
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9999535083770752,
+      "kl_div_avg": 0.026813074946403503,
+      "learning_rate": 4.685314685314685e-07,
+      "loss_func": "stage2",
+      "step": 141,
+      "total_loss": 0.024522602558135986
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9998844265937805,
+      "kl_div_avg": 0.031967416405677795,
+      "learning_rate": 4.682983682983683e-07,
+      "loss_func": "stage2",
+      "step": 142,
+      "total_loss": 0.6268632411956787
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9997915029525757,
+      "kl_div_avg": 0.024857094511389732,
+      "learning_rate": 4.6806526806526804e-07,
+      "loss_func": "stage2",
+      "step": 143,
+      "total_loss": 0.13797396421432495
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9998888969421387,
+      "kl_div_avg": 0.04350988566875458,
+      "learning_rate": 4.6783216783216784e-07,
+      "loss_func": "stage2",
+      "step": 144,
+      "total_loss": 0.5312750935554504
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9998694658279419,
+      "kl_div_avg": 0.03627926483750343,
+      "learning_rate": 4.675990675990676e-07,
+      "loss_func": "stage2",
+      "step": 145,
+      "total_loss": 0.09028466045856476
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9999563694000244,
+      "kl_div_avg": 0.02497359737753868,
+      "learning_rate": 4.673659673659673e-07,
+      "loss_func": "stage2",
+      "step": 146,
+      "total_loss": 0.5303145051002502
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9999274015426636,
+      "kl_div_avg": 0.026778005063533783,
+      "learning_rate": 4.6713286713286707e-07,
+      "loss_func": "stage2",
+      "step": 147,
+      "total_loss": 0.1259535402059555
+    },
+    {
+      "epoch": 0.34,
+      "importance_ratio": 0.9997484683990479,
+      "kl_div_avg": 0.02970227226614952,
+      "learning_rate": 4.668997668997669e-07,
+      "loss_func": "stage2",
+      "step": 148,
+      "total_loss": 0.5203793048858643
+    },
+    {
+      "epoch": 0.35,
+      "importance_ratio": 1.0000306367874146,
+      "kl_div_avg": 0.03883244842290878,
+      "learning_rate": 4.6666666666666666e-07,
+      "loss_func": "stage2",
+      "step": 149,
+      "total_loss": 0.5459209680557251
+    },
+    {
+      "epoch": 0.35,
+      "importance_ratio": 0.9999977946281433,
+      "kl_div_avg": 0.010376233607530594,
+      "learning_rate": 4.664335664335664e-07,
+      "loss_func": "stage2",
+      "step": 150,
+      "total_loss": 0.5408558249473572
+    },
+    {
+      "epoch": 0.35,
+      "importance_ratio": 1.0000672340393066,
+      "kl_div_avg": 0.020404186099767685,
+      "learning_rate": 4.6620046620046615e-07,
+      "loss_func": "stage2",
+      "step": 151,
+      "total_loss": 0.5100921392440796
+    },
+    {
+      "epoch": 0.35,
+      "importance_ratio": 1.0013837814331055,
+      "kl_div_avg": 0.032543182373046875,
+      "learning_rate": 4.6596736596736595e-07,
+      "loss_func": "stage2",
+      "step": 152,
+      "total_loss": -0.4812799394130707
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 2150,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 20.0,
+  "save_steps": 50,
+  "total_flos": 0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:964bcf303f7d37b2140c9e65650e08c3d23d3c46e92f12eb9f174784ccdb4f38
+size 11320

zero_to_fp32.py

@@ -0,0 +1,587 @@
+#!/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 . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+    """
+    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)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+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 _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()
+            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):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    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``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    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_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``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``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+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_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    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("-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_file, tag=args.tag)