add models

mistral_7b_brft/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<pad>": 32000
+}

mistral_7b_brft/config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 32000,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.36.2",
+  "use_cache": false,
+  "vocab_size": 32001
+}

mistral_7b_brft/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.36.2"
+}

mistral_7b_brft/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7df12061950007163481da4da18131d1e6d569fc245d2a7b8baa2e4415a9122
+size 4943170528

mistral_7b_brft/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:581de6ed4df0db8a579756b3f2916caeef99ad5766dddea14414df0aa63dd923
+size 4999819336

mistral_7b_brft/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71c217b611f87450571bf93fc2548df3eb2e7861745bf06fd2a3da0e6df3950f
+size 4540524536

mistral_7b_brft/model.safetensors.index.json

@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 14483480576
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00003-of-00003.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.norm.weight": "model-00003-of-00003.safetensors"
+  }
+}

mistral_7b_brft/special_tokens_map.json

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

mistral_7b_brft/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

mistral_7b_brft/tokenizer_config.json

@@ -0,0 +1,51 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32000": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 2048,
+  "pad_token": "<pad>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

mistral_7b_brft/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:584a12d12d843cfcafa1f1b6387dd7e0891e43cef8f163ccbc36e7947830f760
+size 1976

mistral_7b_step_dpo_var2/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<pad>": 32000
+}

mistral_7b_step_dpo_var2/config.json

@@ -0,0 +1,27 @@
+{
+  "_name_or_path": "/workspace/PRM_PROJECT/models/mistral_7b_brft",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 32000,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.36.2",
+  "use_cache": false,
+  "vocab_size": 32001
+}

mistral_7b_step_dpo_var2/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.36.2"
+}

mistral_7b_step_dpo_var2/latest

@@ -0,0 +1 @@
+global_step1760

mistral_7b_step_dpo_var2/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b50ede466cc29886b510b6323e8c6268aa38e1241834614803a135c7f9a6a17
+size 4943170528

mistral_7b_step_dpo_var2/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6d107fb666c5b6aab09d9ef42e97fc2406e7c49271a5a8a5700431c97b02d91
+size 4999819336

mistral_7b_step_dpo_var2/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:445276552334909131dc57e1a8ecacbd03e21e4efeb79ff9c3b28f0d0e90cc92
+size 4540524536

mistral_7b_step_dpo_var2/model.safetensors.index.json

@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 14483480576
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00003-of-00003.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.norm.weight": "model-00003-of-00003.safetensors"
+  }
+}

mistral_7b_step_dpo_var2/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbe0d720c4c75a6a04213fa3b64bacbe794718a53e2b56ebb67a1a795014dfad
+size 15024

mistral_7b_step_dpo_var2/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72452d3138d0ca2ff89429e3294a834ae7a68e8596fc757735ca56ae52509d57
+size 15024

mistral_7b_step_dpo_var2/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f36e306fb8ebcf53a167bfd6c9af74db410a269ada1e619e3e816f5269543b9d
+size 15024

mistral_7b_step_dpo_var2/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb47ce0c6f815a6f8302b0e3819b4c2315ca71dae3138d97fdceb765cdd0a039
+size 15024

mistral_7b_step_dpo_var2/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ba84746af048747e695b7552d336ad8a7a04cf37e3d7f25d25ef7ebc0193063
+size 1064

mistral_7b_step_dpo_var2/special_tokens_map.json

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

mistral_7b_step_dpo_var2/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
+size 493443

mistral_7b_step_dpo_var2/tokenizer_config.json

@@ -0,0 +1,51 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32000": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 2048,
+  "pad_token": "<pad>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

mistral_7b_step_dpo_var2/trainer_state.json

@@ -0,0 +1,2515 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1760,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.8409090909090907e-09,
+      "logits/chosen": -2.8217740058898926,
+      "logits/rejected": -2.9463119506835938,
+      "logps/chosen": -43.70173263549805,
+      "logps/rejected": -39.51229476928711,
+      "loss": 0.6931,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.0,
+      "rewards/margins": 0.0,
+      "rewards/rejected": 0.0,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.8409090909090908e-08,
+      "logits/chosen": -2.8449532985687256,
+      "logits/rejected": -2.8694777488708496,
+      "logps/chosen": -47.75469207763672,
+      "logps/rejected": -38.914730072021484,
+      "loss": 0.69,
+      "rewards/accuracies": 0.4861111044883728,
+      "rewards/chosen": 0.008462969213724136,
+      "rewards/margins": 0.003949173726141453,
+      "rewards/rejected": 0.0045137968845665455,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.6818181818181815e-08,
+      "logits/chosen": -2.7558584213256836,
+      "logits/rejected": -2.744150400161743,
+      "logps/chosen": -44.27663040161133,
+      "logps/rejected": -37.22570037841797,
+      "loss": 0.6448,
+      "rewards/accuracies": 0.8999999761581421,
+      "rewards/chosen": 0.14198732376098633,
+      "rewards/margins": 0.09488759934902191,
+      "rewards/rejected": 0.04709972068667412,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 8.522727272727271e-08,
+      "logits/chosen": -2.7820916175842285,
+      "logits/rejected": -2.7940614223480225,
+      "logps/chosen": -41.71614456176758,
+      "logps/rejected": -37.8540153503418,
+      "loss": 0.5233,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.6046115159988403,
+      "rewards/margins": 0.4143578112125397,
+      "rewards/rejected": 0.19025376439094543,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1363636363636363e-07,
+      "logits/chosen": -2.7578072547912598,
+      "logits/rejected": -2.7870500087738037,
+      "logps/chosen": -34.32839584350586,
+      "logps/rejected": -36.24866485595703,
+      "loss": 0.3998,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.1275522708892822,
+      "rewards/margins": 0.8746433258056641,
+      "rewards/rejected": 0.25290900468826294,
+      "step": 40
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.4204545454545455e-07,
+      "logits/chosen": -2.7034201622009277,
+      "logits/rejected": -2.7121880054473877,
+      "logps/chosen": -32.550987243652344,
+      "logps/rejected": -36.45380401611328,
+      "loss": 0.2891,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.3929576873779297,
+      "rewards/margins": 1.2349400520324707,
+      "rewards/rejected": 0.15801763534545898,
+      "step": 50
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7045454545454543e-07,
+      "logits/chosen": -2.7104477882385254,
+      "logits/rejected": -2.67173171043396,
+      "logps/chosen": -31.20966148376465,
+      "logps/rejected": -39.04589080810547,
+      "loss": 0.2324,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.6254637241363525,
+      "rewards/margins": 1.5037833452224731,
+      "rewards/rejected": 0.12168042361736298,
+      "step": 60
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9886363636363636e-07,
+      "logits/chosen": -2.6458804607391357,
+      "logits/rejected": -2.6091885566711426,
+      "logps/chosen": -28.253849029541016,
+      "logps/rejected": -39.71971893310547,
+      "loss": 0.1618,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.8234485387802124,
+      "rewards/margins": 1.8900415897369385,
+      "rewards/rejected": -0.06659301370382309,
+      "step": 70
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.2727272727272726e-07,
+      "logits/chosen": -2.6717936992645264,
+      "logits/rejected": -2.6557881832122803,
+      "logps/chosen": -27.09893798828125,
+      "logps/rejected": -42.3707160949707,
+      "loss": 0.1098,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0375189781188965,
+      "rewards/margins": 2.2491657733917236,
+      "rewards/rejected": -0.21164636313915253,
+      "step": 80
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.5568181818181816e-07,
+      "logits/chosen": -2.6540191173553467,
+      "logits/rejected": -2.605602741241455,
+      "logps/chosen": -27.104595184326172,
+      "logps/rejected": -44.61492156982422,
+      "loss": 0.0846,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.16882061958313,
+      "rewards/margins": 2.492210865020752,
+      "rewards/rejected": -0.323390394449234,
+      "step": 90
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2.840909090909091e-07,
+      "logits/chosen": -2.6107547283172607,
+      "logits/rejected": -2.6194005012512207,
+      "logps/chosen": -23.347625732421875,
+      "logps/rejected": -45.462425231933594,
+      "loss": 0.0634,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2358736991882324,
+      "rewards/margins": 2.9278182983398438,
+      "rewards/rejected": -0.6919445991516113,
+      "step": 100
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.1249999999999997e-07,
+      "logits/chosen": -2.595785617828369,
+      "logits/rejected": -2.5354442596435547,
+      "logps/chosen": -23.731903076171875,
+      "logps/rejected": -45.2347297668457,
+      "loss": 0.0524,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2842602729797363,
+      "rewards/margins": 3.1450023651123047,
+      "rewards/rejected": -0.8607422113418579,
+      "step": 110
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.4090909090909085e-07,
+      "logits/chosen": -2.5562920570373535,
+      "logits/rejected": -2.530060291290283,
+      "logps/chosen": -19.341655731201172,
+      "logps/rejected": -50.89426803588867,
+      "loss": 0.0395,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3544158935546875,
+      "rewards/margins": 3.4883244037628174,
+      "rewards/rejected": -1.1339085102081299,
+      "step": 120
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.693181818181818e-07,
+      "logits/chosen": -2.5404000282287598,
+      "logits/rejected": -2.4982523918151855,
+      "logps/chosen": -22.195476531982422,
+      "logps/rejected": -49.90346908569336,
+      "loss": 0.0339,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.340710401535034,
+      "rewards/margins": 3.5701756477355957,
+      "rewards/rejected": -1.2294648885726929,
+      "step": 130
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 3.977272727272727e-07,
+      "logits/chosen": -2.560086965560913,
+      "logits/rejected": -2.4935762882232666,
+      "logps/chosen": -23.410430908203125,
+      "logps/rejected": -53.67076873779297,
+      "loss": 0.0303,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.279090404510498,
+      "rewards/margins": 3.8115992546081543,
+      "rewards/rejected": -1.5325093269348145,
+      "step": 140
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.2613636363636364e-07,
+      "logits/chosen": -2.559985637664795,
+      "logits/rejected": -2.504645824432373,
+      "logps/chosen": -26.432170867919922,
+      "logps/rejected": -57.561363220214844,
+      "loss": 0.0226,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.357306957244873,
+      "rewards/margins": 4.020358085632324,
+      "rewards/rejected": -1.6630510091781616,
+      "step": 150
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.545454545454545e-07,
+      "logits/chosen": -2.567091464996338,
+      "logits/rejected": -2.4482944011688232,
+      "logps/chosen": -23.907033920288086,
+      "logps/rejected": -57.371864318847656,
+      "loss": 0.0181,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3331665992736816,
+      "rewards/margins": 4.317889213562012,
+      "rewards/rejected": -1.984722375869751,
+      "step": 160
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.829545454545455e-07,
+      "logits/chosen": -2.470553159713745,
+      "logits/rejected": -2.4206271171569824,
+      "logps/chosen": -23.748369216918945,
+      "logps/rejected": -64.3258056640625,
+      "loss": 0.0151,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3254458904266357,
+      "rewards/margins": 4.598530292510986,
+      "rewards/rejected": -2.2730841636657715,
+      "step": 170
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.987373737373738e-07,
+      "logits/chosen": -2.4871647357940674,
+      "logits/rejected": -2.396991014480591,
+      "logps/chosen": -27.013507843017578,
+      "logps/rejected": -62.3752326965332,
+      "loss": 0.0148,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3036725521087646,
+      "rewards/margins": 4.721800804138184,
+      "rewards/rejected": -2.418128252029419,
+      "step": 180
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.95580808080808e-07,
+      "logits/chosen": -2.4113528728485107,
+      "logits/rejected": -2.3845059871673584,
+      "logps/chosen": -23.671329498291016,
+      "logps/rejected": -62.555519104003906,
+      "loss": 0.0122,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1994900703430176,
+      "rewards/margins": 4.782956123352051,
+      "rewards/rejected": -2.583466053009033,
+      "step": 190
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.924242424242424e-07,
+      "logits/chosen": -2.435971975326538,
+      "logits/rejected": -2.287334680557251,
+      "logps/chosen": -26.926639556884766,
+      "logps/rejected": -65.13796997070312,
+      "loss": 0.0119,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1797242164611816,
+      "rewards/margins": 4.992079734802246,
+      "rewards/rejected": -2.8123557567596436,
+      "step": 200
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.892676767676767e-07,
+      "logits/chosen": -2.3854382038116455,
+      "logits/rejected": -2.3133955001831055,
+      "logps/chosen": -22.9410400390625,
+      "logps/rejected": -70.91532897949219,
+      "loss": 0.0078,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2163994312286377,
+      "rewards/margins": 5.3300251960754395,
+      "rewards/rejected": -3.113626003265381,
+      "step": 210
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.861111111111111e-07,
+      "logits/chosen": -2.3645615577697754,
+      "logits/rejected": -2.1975860595703125,
+      "logps/chosen": -23.414844512939453,
+      "logps/rejected": -70.23318481445312,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1890006065368652,
+      "rewards/margins": 5.3253374099731445,
+      "rewards/rejected": -3.136336326599121,
+      "step": 220
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.829545454545455e-07,
+      "logits/chosen": -2.364175319671631,
+      "logits/rejected": -2.2675890922546387,
+      "logps/chosen": -25.140579223632812,
+      "logps/rejected": -70.91351318359375,
+      "loss": 0.0064,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1673789024353027,
+      "rewards/margins": 5.609480857849121,
+      "rewards/rejected": -3.4421024322509766,
+      "step": 230
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.797979797979798e-07,
+      "logits/chosen": -2.3288331031799316,
+      "logits/rejected": -2.1772916316986084,
+      "logps/chosen": -23.989219665527344,
+      "logps/rejected": -77.8015365600586,
+      "loss": 0.006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3028969764709473,
+      "rewards/margins": 5.861783981323242,
+      "rewards/rejected": -3.558886766433716,
+      "step": 240
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.766414141414141e-07,
+      "logits/chosen": -2.240269184112549,
+      "logits/rejected": -2.064934492111206,
+      "logps/chosen": -23.393871307373047,
+      "logps/rejected": -72.64311218261719,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1841888427734375,
+      "rewards/margins": 5.726863861083984,
+      "rewards/rejected": -3.5426745414733887,
+      "step": 250
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.734848484848485e-07,
+      "logits/chosen": -2.2834770679473877,
+      "logits/rejected": -2.0509936809539795,
+      "logps/chosen": -23.929466247558594,
+      "logps/rejected": -74.3910903930664,
+      "loss": 0.0053,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2868974208831787,
+      "rewards/margins": 5.918667793273926,
+      "rewards/rejected": -3.631770372390747,
+      "step": 260
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.703282828282828e-07,
+      "logits/chosen": -2.264416217803955,
+      "logits/rejected": -1.8542054891586304,
+      "logps/chosen": -24.270587921142578,
+      "logps/rejected": -79.36566162109375,
+      "loss": 0.0035,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2542290687561035,
+      "rewards/margins": 6.15808629989624,
+      "rewards/rejected": -3.903857707977295,
+      "step": 270
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.6717171717171714e-07,
+      "logits/chosen": -2.232429265975952,
+      "logits/rejected": -2.002338409423828,
+      "logps/chosen": -23.89547348022461,
+      "logps/rejected": -78.91621398925781,
+      "loss": 0.0033,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1463963985443115,
+      "rewards/margins": 6.129814624786377,
+      "rewards/rejected": -3.9834182262420654,
+      "step": 280
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.640151515151515e-07,
+      "logits/chosen": -2.2267403602600098,
+      "logits/rejected": -1.9576308727264404,
+      "logps/chosen": -22.615476608276367,
+      "logps/rejected": -82.0691909790039,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2505202293395996,
+      "rewards/margins": 6.612841606140137,
+      "rewards/rejected": -4.3623223304748535,
+      "step": 290
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.6085858585858586e-07,
+      "logits/chosen": -2.1656718254089355,
+      "logits/rejected": -1.964550256729126,
+      "logps/chosen": -27.56490707397461,
+      "logps/rejected": -81.64122009277344,
+      "loss": 0.0054,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.032174825668335,
+      "rewards/margins": 6.209091663360596,
+      "rewards/rejected": -4.176916599273682,
+      "step": 300
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.577020202020202e-07,
+      "logits/chosen": -2.135385274887085,
+      "logits/rejected": -1.9940789937973022,
+      "logps/chosen": -25.354991912841797,
+      "logps/rejected": -83.73381042480469,
+      "loss": 0.0027,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1884491443634033,
+      "rewards/margins": 6.562999725341797,
+      "rewards/rejected": -4.3745503425598145,
+      "step": 310
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.545454545454545e-07,
+      "logits/chosen": -2.1604058742523193,
+      "logits/rejected": -1.9300079345703125,
+      "logps/chosen": -25.576618194580078,
+      "logps/rejected": -81.22620391845703,
+      "loss": 0.0021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2594103813171387,
+      "rewards/margins": 6.653275966644287,
+      "rewards/rejected": -4.393866062164307,
+      "step": 320
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.513888888888889e-07,
+      "logits/chosen": -2.092848062515259,
+      "logits/rejected": -1.8601881265640259,
+      "logps/chosen": -30.18955421447754,
+      "logps/rejected": -86.6014175415039,
+      "loss": 0.0024,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.9979768991470337,
+      "rewards/margins": 6.703522682189941,
+      "rewards/rejected": -4.705544471740723,
+      "step": 330
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.4823232323232324e-07,
+      "logits/chosen": -2.140770196914673,
+      "logits/rejected": -1.8940891027450562,
+      "logps/chosen": -24.722496032714844,
+      "logps/rejected": -87.34082794189453,
+      "loss": 0.0032,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.163637638092041,
+      "rewards/margins": 6.7660322189331055,
+      "rewards/rejected": -4.602395057678223,
+      "step": 340
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.4507575757575754e-07,
+      "logits/chosen": -2.143838405609131,
+      "logits/rejected": -1.9490582942962646,
+      "logps/chosen": -25.524768829345703,
+      "logps/rejected": -87.40507507324219,
+      "loss": 0.002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1567718982696533,
+      "rewards/margins": 6.852728366851807,
+      "rewards/rejected": -4.695956230163574,
+      "step": 350
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.419191919191919e-07,
+      "logits/chosen": -2.1022706031799316,
+      "logits/rejected": -1.7483265399932861,
+      "logps/chosen": -25.361112594604492,
+      "logps/rejected": -85.70404052734375,
+      "loss": 0.0019,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.186305522918701,
+      "rewards/margins": 6.8868727684021,
+      "rewards/rejected": -4.70056676864624,
+      "step": 360
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.387626262626262e-07,
+      "logits/chosen": -2.1865038871765137,
+      "logits/rejected": -1.818566083908081,
+      "logps/chosen": -24.06246566772461,
+      "logps/rejected": -87.37457275390625,
+      "loss": 0.0014,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2870378494262695,
+      "rewards/margins": 7.144204139709473,
+      "rewards/rejected": -4.857166767120361,
+      "step": 370
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.3560606060606057e-07,
+      "logits/chosen": -2.1236586570739746,
+      "logits/rejected": -1.8356149196624756,
+      "logps/chosen": -26.733007431030273,
+      "logps/rejected": -86.30492401123047,
+      "loss": 0.0014,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.197669267654419,
+      "rewards/margins": 6.9382452964782715,
+      "rewards/rejected": -4.740575790405273,
+      "step": 380
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.32449494949495e-07,
+      "logits/chosen": -2.069997787475586,
+      "logits/rejected": -1.772733449935913,
+      "logps/chosen": -23.993722915649414,
+      "logps/rejected": -88.73332977294922,
+      "loss": 0.0018,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.214871883392334,
+      "rewards/margins": 7.181988716125488,
+      "rewards/rejected": -4.967116355895996,
+      "step": 390
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.292929292929293e-07,
+      "logits/chosen": -2.0845890045166016,
+      "logits/rejected": -1.7745990753173828,
+      "logps/chosen": -27.915542602539062,
+      "logps/rejected": -85.68502044677734,
+      "loss": 0.0028,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.107525587081909,
+      "rewards/margins": 6.8236236572265625,
+      "rewards/rejected": -4.716097831726074,
+      "step": 400
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.2613636363636364e-07,
+      "logits/chosen": -2.119929790496826,
+      "logits/rejected": -1.7703702449798584,
+      "logps/chosen": -26.399669647216797,
+      "logps/rejected": -91.058349609375,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.180525779724121,
+      "rewards/margins": 7.306072235107422,
+      "rewards/rejected": -5.125546455383301,
+      "step": 410
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.2297979797979795e-07,
+      "logits/chosen": -2.053358793258667,
+      "logits/rejected": -1.8158977031707764,
+      "logps/chosen": -25.54621696472168,
+      "logps/rejected": -88.72142028808594,
+      "loss": 0.0016,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2602436542510986,
+      "rewards/margins": 7.221992492675781,
+      "rewards/rejected": -4.961750030517578,
+      "step": 420
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.198232323232323e-07,
+      "logits/chosen": -2.0827341079711914,
+      "logits/rejected": -1.6577997207641602,
+      "logps/chosen": -27.02775001525879,
+      "logps/rejected": -91.33396911621094,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1475205421447754,
+      "rewards/margins": 7.185147762298584,
+      "rewards/rejected": -5.037627220153809,
+      "step": 430
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.1666666666666667e-07,
+      "logits/chosen": -2.0534794330596924,
+      "logits/rejected": -1.7211973667144775,
+      "logps/chosen": -23.891704559326172,
+      "logps/rejected": -91.61319732666016,
+      "loss": 0.0014,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.24546217918396,
+      "rewards/margins": 7.388940334320068,
+      "rewards/rejected": -5.143477916717529,
+      "step": 440
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.1351010101010097e-07,
+      "logits/chosen": -2.0260019302368164,
+      "logits/rejected": -1.7830889225006104,
+      "logps/chosen": -26.86496925354004,
+      "logps/rejected": -92.28340148925781,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.155402660369873,
+      "rewards/margins": 7.439195156097412,
+      "rewards/rejected": -5.283791542053223,
+      "step": 450
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.1035353535353533e-07,
+      "logits/chosen": -2.027803659439087,
+      "logits/rejected": -1.83200204372406,
+      "logps/chosen": -23.117488861083984,
+      "logps/rejected": -87.80825805664062,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.3013126850128174,
+      "rewards/margins": 7.291370391845703,
+      "rewards/rejected": -4.990058898925781,
+      "step": 460
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.071969696969697e-07,
+      "logits/chosen": -1.9879964590072632,
+      "logits/rejected": -1.808397889137268,
+      "logps/chosen": -24.546789169311523,
+      "logps/rejected": -92.23595428466797,
+      "loss": 0.0026,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.160706043243408,
+      "rewards/margins": 7.311809539794922,
+      "rewards/rejected": -5.1511030197143555,
+      "step": 470
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.04040404040404e-07,
+      "logits/chosen": -1.9482005834579468,
+      "logits/rejected": -1.6757386922836304,
+      "logps/chosen": -24.567623138427734,
+      "logps/rejected": -90.96485900878906,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0390470027923584,
+      "rewards/margins": 7.19284725189209,
+      "rewards/rejected": -5.1538004875183105,
+      "step": 480
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.008838383838384e-07,
+      "logits/chosen": -2.0609028339385986,
+      "logits/rejected": -1.8258024454116821,
+      "logps/chosen": -27.727807998657227,
+      "logps/rejected": -90.19026947021484,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1395559310913086,
+      "rewards/margins": 7.3105363845825195,
+      "rewards/rejected": -5.170979976654053,
+      "step": 490
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.977272727272727e-07,
+      "logits/chosen": -1.9994752407073975,
+      "logits/rejected": -1.7483898401260376,
+      "logps/chosen": -23.689210891723633,
+      "logps/rejected": -90.74690246582031,
+      "loss": 0.0011,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.228365182876587,
+      "rewards/margins": 7.52515172958374,
+      "rewards/rejected": -5.296786308288574,
+      "step": 500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.9457070707070707e-07,
+      "logits/chosen": -1.9040803909301758,
+      "logits/rejected": -1.6641581058502197,
+      "logps/chosen": -26.4534969329834,
+      "logps/rejected": -90.5098876953125,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.180100202560425,
+      "rewards/margins": 7.4137864112854,
+      "rewards/rejected": -5.233686447143555,
+      "step": 510
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.9141414141414143e-07,
+      "logits/chosen": -1.992407202720642,
+      "logits/rejected": -1.7667903900146484,
+      "logps/chosen": -26.068599700927734,
+      "logps/rejected": -90.90068054199219,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.093018054962158,
+      "rewards/margins": 7.419617652893066,
+      "rewards/rejected": -5.326600074768066,
+      "step": 520
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.8825757575757573e-07,
+      "logits/chosen": -1.976304292678833,
+      "logits/rejected": -1.7016494274139404,
+      "logps/chosen": -25.243732452392578,
+      "logps/rejected": -91.59854125976562,
+      "loss": 0.0023,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.261107921600342,
+      "rewards/margins": 7.474202632904053,
+      "rewards/rejected": -5.213095188140869,
+      "step": 530
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.851010101010101e-07,
+      "logits/chosen": -1.9485681056976318,
+      "logits/rejected": -1.614693284034729,
+      "logps/chosen": -23.00712013244629,
+      "logps/rejected": -93.22959899902344,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1281652450561523,
+      "rewards/margins": 7.445710182189941,
+      "rewards/rejected": -5.317544460296631,
+      "step": 540
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.819444444444444e-07,
+      "logits/chosen": -1.9920011758804321,
+      "logits/rejected": -1.6840722560882568,
+      "logps/chosen": -26.918170928955078,
+      "logps/rejected": -92.25691986083984,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1348118782043457,
+      "rewards/margins": 7.502300262451172,
+      "rewards/rejected": -5.367488861083984,
+      "step": 550
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.7878787878787876e-07,
+      "logits/chosen": -1.9384076595306396,
+      "logits/rejected": -1.7479146718978882,
+      "logps/chosen": -28.570159912109375,
+      "logps/rejected": -92.83436584472656,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.056058168411255,
+      "rewards/margins": 7.413699150085449,
+      "rewards/rejected": -5.357641696929932,
+      "step": 560
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.756313131313131e-07,
+      "logits/chosen": -1.9707496166229248,
+      "logits/rejected": -1.6780744791030884,
+      "logps/chosen": -23.548856735229492,
+      "logps/rejected": -97.84285736083984,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.162672519683838,
+      "rewards/margins": 7.8429388999938965,
+      "rewards/rejected": -5.6802659034729,
+      "step": 570
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.724747474747474e-07,
+      "logits/chosen": -1.9848487377166748,
+      "logits/rejected": -1.7538635730743408,
+      "logps/chosen": -26.496463775634766,
+      "logps/rejected": -91.71202087402344,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.149207592010498,
+      "rewards/margins": 7.3838911056518555,
+      "rewards/rejected": -5.234683990478516,
+      "step": 580
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 3.693181818181818e-07,
+      "logits/chosen": -2.0094056129455566,
+      "logits/rejected": -1.6735060214996338,
+      "logps/chosen": -27.791439056396484,
+      "logps/rejected": -93.74909973144531,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.057635545730591,
+      "rewards/margins": 7.495272159576416,
+      "rewards/rejected": -5.437636852264404,
+      "step": 590
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 3.661616161616162e-07,
+      "logits/chosen": -1.9600168466567993,
+      "logits/rejected": -1.7584631443023682,
+      "logps/chosen": -25.7122859954834,
+      "logps/rejected": -94.01753997802734,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.14912486076355,
+      "rewards/margins": 7.507033348083496,
+      "rewards/rejected": -5.357908248901367,
+      "step": 600
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 3.630050505050505e-07,
+      "logits/chosen": -2.021538257598877,
+      "logits/rejected": -1.6666122674942017,
+      "logps/chosen": -23.989110946655273,
+      "logps/rejected": -97.35430145263672,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2275078296661377,
+      "rewards/margins": 7.870290279388428,
+      "rewards/rejected": -5.642783164978027,
+      "step": 610
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 3.5984848484848486e-07,
+      "logits/chosen": -1.8722727298736572,
+      "logits/rejected": -1.661717176437378,
+      "logps/chosen": -23.958885192871094,
+      "logps/rejected": -95.70252227783203,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1420698165893555,
+      "rewards/margins": 7.835183143615723,
+      "rewards/rejected": -5.693112850189209,
+      "step": 620
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 3.5669191919191916e-07,
+      "logits/chosen": -2.0347046852111816,
+      "logits/rejected": -1.5893232822418213,
+      "logps/chosen": -26.783245086669922,
+      "logps/rejected": -93.17195892333984,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1823782920837402,
+      "rewards/margins": 7.69185733795166,
+      "rewards/rejected": -5.509478569030762,
+      "step": 630
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 3.535353535353535e-07,
+      "logits/chosen": -1.925437569618225,
+      "logits/rejected": -1.5153725147247314,
+      "logps/chosen": -23.424793243408203,
+      "logps/rejected": -96.51454162597656,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2793221473693848,
+      "rewards/margins": 7.9029412269592285,
+      "rewards/rejected": -5.62362003326416,
+      "step": 640
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 3.503787878787879e-07,
+      "logits/chosen": -1.9150264263153076,
+      "logits/rejected": -1.7848446369171143,
+      "logps/chosen": -26.177997589111328,
+      "logps/rejected": -93.8024673461914,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1856698989868164,
+      "rewards/margins": 7.7468581199646,
+      "rewards/rejected": -5.561188697814941,
+      "step": 650
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 3.472222222222222e-07,
+      "logits/chosen": -1.958645224571228,
+      "logits/rejected": -1.6190398931503296,
+      "logps/chosen": -21.174949645996094,
+      "logps/rejected": -92.42906951904297,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.322924852371216,
+      "rewards/margins": 7.8906145095825195,
+      "rewards/rejected": -5.567690372467041,
+      "step": 660
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 3.4406565656565654e-07,
+      "logits/chosen": -1.9614722728729248,
+      "logits/rejected": -1.5498616695404053,
+      "logps/chosen": -27.733715057373047,
+      "logps/rejected": -96.60962677001953,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.133537769317627,
+      "rewards/margins": 7.692742347717285,
+      "rewards/rejected": -5.559204578399658,
+      "step": 670
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 3.4090909090909085e-07,
+      "logits/chosen": -1.8956854343414307,
+      "logits/rejected": -1.5820499658584595,
+      "logps/chosen": -23.548494338989258,
+      "logps/rejected": -93.71937561035156,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.226599931716919,
+      "rewards/margins": 7.865328311920166,
+      "rewards/rejected": -5.63872766494751,
+      "step": 680
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 3.377525252525252e-07,
+      "logits/chosen": -1.8554881811141968,
+      "logits/rejected": -1.5859127044677734,
+      "logps/chosen": -24.2250919342041,
+      "logps/rejected": -92.75028991699219,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1412949562072754,
+      "rewards/margins": 7.787182807922363,
+      "rewards/rejected": -5.645887851715088,
+      "step": 690
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 3.345959595959596e-07,
+      "logits/chosen": -1.992211937904358,
+      "logits/rejected": -1.420165777206421,
+      "logps/chosen": -28.7690372467041,
+      "logps/rejected": -97.75999450683594,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.145841598510742,
+      "rewards/margins": 7.845877647399902,
+      "rewards/rejected": -5.70003604888916,
+      "step": 700
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 3.314393939393939e-07,
+      "logits/chosen": -1.8932504653930664,
+      "logits/rejected": -1.5480964183807373,
+      "logps/chosen": -24.157947540283203,
+      "logps/rejected": -92.73865509033203,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1332340240478516,
+      "rewards/margins": 7.6235527992248535,
+      "rewards/rejected": -5.490318298339844,
+      "step": 710
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 3.282828282828283e-07,
+      "logits/chosen": -1.8008241653442383,
+      "logits/rejected": -1.527385950088501,
+      "logps/chosen": -23.674373626708984,
+      "logps/rejected": -96.29564666748047,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.252340316772461,
+      "rewards/margins": 8.004530906677246,
+      "rewards/rejected": -5.752191066741943,
+      "step": 720
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 3.2512626262626264e-07,
+      "logits/chosen": -1.9497926235198975,
+      "logits/rejected": -1.363884687423706,
+      "logps/chosen": -23.689098358154297,
+      "logps/rejected": -94.29708862304688,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0340566635131836,
+      "rewards/margins": 7.854531764984131,
+      "rewards/rejected": -5.820474624633789,
+      "step": 730
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.2196969696969695e-07,
+      "logits/chosen": -1.8968747854232788,
+      "logits/rejected": -1.424194097518921,
+      "logps/chosen": -26.381500244140625,
+      "logps/rejected": -99.07059478759766,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1705589294433594,
+      "rewards/margins": 8.047155380249023,
+      "rewards/rejected": -5.876595497131348,
+      "step": 740
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.188131313131313e-07,
+      "logits/chosen": -1.7851377725601196,
+      "logits/rejected": -1.5909427404403687,
+      "logps/chosen": -26.67759132385254,
+      "logps/rejected": -95.57559967041016,
+      "loss": 0.0013,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.01067852973938,
+      "rewards/margins": 7.764972686767578,
+      "rewards/rejected": -5.754294395446777,
+      "step": 750
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.156565656565656e-07,
+      "logits/chosen": -2.0088281631469727,
+      "logits/rejected": -1.432469129562378,
+      "logps/chosen": -25.423076629638672,
+      "logps/rejected": -96.91383361816406,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.265986919403076,
+      "rewards/margins": 8.044832229614258,
+      "rewards/rejected": -5.778844833374023,
+      "step": 760
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.1249999999999997e-07,
+      "logits/chosen": -1.971697211265564,
+      "logits/rejected": -1.3336002826690674,
+      "logps/chosen": -27.496530532836914,
+      "logps/rejected": -99.48794555664062,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1343042850494385,
+      "rewards/margins": 8.016721725463867,
+      "rewards/rejected": -5.88241720199585,
+      "step": 770
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.0934343434343433e-07,
+      "logits/chosen": -1.9239240884780884,
+      "logits/rejected": -1.4457045793533325,
+      "logps/chosen": -24.486019134521484,
+      "logps/rejected": -99.0895767211914,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1495842933654785,
+      "rewards/margins": 8.156563758850098,
+      "rewards/rejected": -6.006979942321777,
+      "step": 780
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.0618686868686864e-07,
+      "logits/chosen": -1.9241794347763062,
+      "logits/rejected": -1.5538829565048218,
+      "logps/chosen": -23.10116958618164,
+      "logps/rejected": -95.69884490966797,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.230299472808838,
+      "rewards/margins": 8.168355941772461,
+      "rewards/rejected": -5.938056945800781,
+      "step": 790
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.0303030303030305e-07,
+      "logits/chosen": -1.8651403188705444,
+      "logits/rejected": -1.6048393249511719,
+      "logps/chosen": -24.562875747680664,
+      "logps/rejected": -100.53028106689453,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.23217511177063,
+      "rewards/margins": 8.34048843383789,
+      "rewards/rejected": -6.10831356048584,
+      "step": 800
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 2.9987373737373735e-07,
+      "logits/chosen": -1.9406903982162476,
+      "logits/rejected": -1.516804814338684,
+      "logps/chosen": -27.88791847229004,
+      "logps/rejected": -99.57067108154297,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.297253131866455,
+      "rewards/margins": 8.173368453979492,
+      "rewards/rejected": -5.8761162757873535,
+      "step": 810
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.967171717171717e-07,
+      "logits/chosen": -1.8654781579971313,
+      "logits/rejected": -1.4960870742797852,
+      "logps/chosen": -26.56595230102539,
+      "logps/rejected": -96.1545181274414,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.13999605178833,
+      "rewards/margins": 8.020753860473633,
+      "rewards/rejected": -5.880757808685303,
+      "step": 820
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 2.9356060606060607e-07,
+      "logits/chosen": -1.7485897541046143,
+      "logits/rejected": -1.216596245765686,
+      "logps/chosen": -29.345123291015625,
+      "logps/rejected": -96.46388244628906,
+      "loss": 0.0021,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.930932641029358,
+      "rewards/margins": 7.820921421051025,
+      "rewards/rejected": -5.889988899230957,
+      "step": 830
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.904040404040404e-07,
+      "logits/chosen": -1.7588354349136353,
+      "logits/rejected": -1.245519995689392,
+      "logps/chosen": -25.377761840820312,
+      "logps/rejected": -97.64041900634766,
+      "loss": 0.0011,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.127272129058838,
+      "rewards/margins": 7.99755334854126,
+      "rewards/rejected": -5.87028169631958,
+      "step": 840
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 2.8724747474747474e-07,
+      "logits/chosen": -1.7543283700942993,
+      "logits/rejected": -1.163822889328003,
+      "logps/chosen": -25.090377807617188,
+      "logps/rejected": -97.6384048461914,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2114098072052,
+      "rewards/margins": 8.035085678100586,
+      "rewards/rejected": -5.823675632476807,
+      "step": 850
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.840909090909091e-07,
+      "logits/chosen": -1.8108580112457275,
+      "logits/rejected": -1.1935783624649048,
+      "logps/chosen": -25.921573638916016,
+      "logps/rejected": -96.55907440185547,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.208132266998291,
+      "rewards/margins": 8.017982482910156,
+      "rewards/rejected": -5.809850215911865,
+      "step": 860
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 2.809343434343434e-07,
+      "logits/chosen": -1.7801116704940796,
+      "logits/rejected": -1.4260514974594116,
+      "logps/chosen": -27.629343032836914,
+      "logps/rejected": -96.98956298828125,
+      "loss": 0.008,
+      "rewards/accuracies": 0.987500011920929,
+      "rewards/chosen": 2.035959482192993,
+      "rewards/margins": 7.815830230712891,
+      "rewards/rejected": -5.77987003326416,
+      "step": 870
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 2.7777777777777776e-07,
+      "logits/chosen": -2.078132152557373,
+      "logits/rejected": -1.6956380605697632,
+      "logps/chosen": -25.696014404296875,
+      "logps/rejected": -95.23052978515625,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1760764122009277,
+      "rewards/margins": 7.6634016036987305,
+      "rewards/rejected": -5.487325191497803,
+      "step": 880
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.7462121212121207e-07,
+      "logits/chosen": -2.1057019233703613,
+      "logits/rejected": -1.7346597909927368,
+      "logps/chosen": -27.921838760375977,
+      "logps/rejected": -93.75882720947266,
+      "loss": 0.0018,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.119664192199707,
+      "rewards/margins": 7.701469421386719,
+      "rewards/rejected": -5.581804275512695,
+      "step": 890
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 2.714646464646465e-07,
+      "logits/chosen": -2.072180986404419,
+      "logits/rejected": -1.5586504936218262,
+      "logps/chosen": -26.299549102783203,
+      "logps/rejected": -96.15899658203125,
+      "loss": 0.0015,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1264266967773438,
+      "rewards/margins": 7.7318243980407715,
+      "rewards/rejected": -5.605398178100586,
+      "step": 900
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.6830808080808084e-07,
+      "logits/chosen": -2.08247709274292,
+      "logits/rejected": -1.6104835271835327,
+      "logps/chosen": -25.7777156829834,
+      "logps/rejected": -95.84490966796875,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0905160903930664,
+      "rewards/margins": 7.827942848205566,
+      "rewards/rejected": -5.737426280975342,
+      "step": 910
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.6515151515151514e-07,
+      "logits/chosen": -2.041508674621582,
+      "logits/rejected": -1.6706491708755493,
+      "logps/chosen": -24.13426399230957,
+      "logps/rejected": -94.98015594482422,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1071910858154297,
+      "rewards/margins": 7.764508247375488,
+      "rewards/rejected": -5.6573166847229,
+      "step": 920
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.619949494949495e-07,
+      "logits/chosen": -2.016854763031006,
+      "logits/rejected": -1.7118686437606812,
+      "logps/chosen": -26.060617446899414,
+      "logps/rejected": -95.73835754394531,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.112154006958008,
+      "rewards/margins": 7.850780487060547,
+      "rewards/rejected": -5.738626956939697,
+      "step": 930
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.588383838383838e-07,
+      "logits/chosen": -2.100552797317505,
+      "logits/rejected": -1.6788053512573242,
+      "logps/chosen": -29.341171264648438,
+      "logps/rejected": -96.46426391601562,
+      "loss": 0.002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.104468822479248,
+      "rewards/margins": 7.92138671875,
+      "rewards/rejected": -5.816919326782227,
+      "step": 940
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.5568181818181816e-07,
+      "logits/chosen": -2.0833117961883545,
+      "logits/rejected": -1.65346360206604,
+      "logps/chosen": -25.34517478942871,
+      "logps/rejected": -98.56163787841797,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1046013832092285,
+      "rewards/margins": 7.977512359619141,
+      "rewards/rejected": -5.872910976409912,
+      "step": 950
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.525252525252525e-07,
+      "logits/chosen": -2.103358745574951,
+      "logits/rejected": -1.624432921409607,
+      "logps/chosen": -26.02530860900879,
+      "logps/rejected": -92.7767562866211,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.214033842086792,
+      "rewards/margins": 7.996707916259766,
+      "rewards/rejected": -5.782674312591553,
+      "step": 960
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.493686868686869e-07,
+      "logits/chosen": -2.082728147506714,
+      "logits/rejected": -1.4827877283096313,
+      "logps/chosen": -24.34432601928711,
+      "logps/rejected": -99.1805191040039,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.210329532623291,
+      "rewards/margins": 8.068390846252441,
+      "rewards/rejected": -5.858060836791992,
+      "step": 970
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.462121212121212e-07,
+      "logits/chosen": -2.1130552291870117,
+      "logits/rejected": -1.6159216165542603,
+      "logps/chosen": -26.011852264404297,
+      "logps/rejected": -99.8711166381836,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.144451379776001,
+      "rewards/margins": 8.234131813049316,
+      "rewards/rejected": -6.08967924118042,
+      "step": 980
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.4305555555555555e-07,
+      "logits/chosen": -2.0969700813293457,
+      "logits/rejected": -1.633567452430725,
+      "logps/chosen": -26.281970977783203,
+      "logps/rejected": -98.82225036621094,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.164041757583618,
+      "rewards/margins": 8.042884826660156,
+      "rewards/rejected": -5.878843784332275,
+      "step": 990
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.398989898989899e-07,
+      "logits/chosen": -2.05826473236084,
+      "logits/rejected": -1.5802359580993652,
+      "logps/chosen": -27.45217514038086,
+      "logps/rejected": -94.94364166259766,
+      "loss": 0.0014,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.146556854248047,
+      "rewards/margins": 7.836559295654297,
+      "rewards/rejected": -5.690001964569092,
+      "step": 1000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.3674242424242424e-07,
+      "logits/chosen": -2.138557195663452,
+      "logits/rejected": -1.5822288990020752,
+      "logps/chosen": -24.474956512451172,
+      "logps/rejected": -99.55738830566406,
+      "loss": 0.0012,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1268579959869385,
+      "rewards/margins": 8.155145645141602,
+      "rewards/rejected": -6.028287410736084,
+      "step": 1010
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.3358585858585857e-07,
+      "logits/chosen": -2.0809974670410156,
+      "logits/rejected": -1.586019515991211,
+      "logps/chosen": -25.225772857666016,
+      "logps/rejected": -97.1346435546875,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.189413547515869,
+      "rewards/margins": 8.057894706726074,
+      "rewards/rejected": -5.868481159210205,
+      "step": 1020
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3042929292929293e-07,
+      "logits/chosen": -2.0678093433380127,
+      "logits/rejected": -1.527854323387146,
+      "logps/chosen": -27.391677856445312,
+      "logps/rejected": -99.27685546875,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.080334186553955,
+      "rewards/margins": 8.150042533874512,
+      "rewards/rejected": -6.069709300994873,
+      "step": 1030
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.2727272727272726e-07,
+      "logits/chosen": -2.0118675231933594,
+      "logits/rejected": -1.3917840719223022,
+      "logps/chosen": -25.736526489257812,
+      "logps/rejected": -100.50688171386719,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0019564628601074,
+      "rewards/margins": 7.99700927734375,
+      "rewards/rejected": -5.995053291320801,
+      "step": 1040
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2411616161616162e-07,
+      "logits/chosen": -1.9202121496200562,
+      "logits/rejected": -1.4258867502212524,
+      "logps/chosen": -24.735958099365234,
+      "logps/rejected": -99.00466918945312,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0488181114196777,
+      "rewards/margins": 8.070080757141113,
+      "rewards/rejected": -6.021262168884277,
+      "step": 1050
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2095959595959595e-07,
+      "logits/chosen": -2.0327675342559814,
+      "logits/rejected": -1.523757815361023,
+      "logps/chosen": -24.49960708618164,
+      "logps/rejected": -97.82597351074219,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.147500514984131,
+      "rewards/margins": 8.13215446472168,
+      "rewards/rejected": -5.984654903411865,
+      "step": 1060
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1780303030303028e-07,
+      "logits/chosen": -2.0616095066070557,
+      "logits/rejected": -1.5044522285461426,
+      "logps/chosen": -30.229473114013672,
+      "logps/rejected": -100.33226013183594,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0959255695343018,
+      "rewards/margins": 8.124457359313965,
+      "rewards/rejected": -6.028532981872559,
+      "step": 1070
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1464646464646464e-07,
+      "logits/chosen": -1.9552123546600342,
+      "logits/rejected": -1.5102477073669434,
+      "logps/chosen": -27.121322631835938,
+      "logps/rejected": -100.54606628417969,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 1.937917709350586,
+      "rewards/margins": 8.002965927124023,
+      "rewards/rejected": -6.0650482177734375,
+      "step": 1080
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.1148989898989897e-07,
+      "logits/chosen": -2.0241737365722656,
+      "logits/rejected": -1.5654652118682861,
+      "logps/chosen": -26.224910736083984,
+      "logps/rejected": -97.28495788574219,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.198258876800537,
+      "rewards/margins": 8.065521240234375,
+      "rewards/rejected": -5.867262840270996,
+      "step": 1090
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0833333333333333e-07,
+      "logits/chosen": -1.9633772373199463,
+      "logits/rejected": -1.4830753803253174,
+      "logps/chosen": -26.038715362548828,
+      "logps/rejected": -97.33109283447266,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.152127504348755,
+      "rewards/margins": 8.060604095458984,
+      "rewards/rejected": -5.908476829528809,
+      "step": 1100
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.0517676767676766e-07,
+      "logits/chosen": -1.9758943319320679,
+      "logits/rejected": -1.5877580642700195,
+      "logps/chosen": -26.84524917602539,
+      "logps/rejected": -100.64767456054688,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.031940460205078,
+      "rewards/margins": 7.97409200668335,
+      "rewards/rejected": -5.942150592803955,
+      "step": 1110
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 2.02020202020202e-07,
+      "logits/chosen": -1.9573463201522827,
+      "logits/rejected": -1.6327768564224243,
+      "logps/chosen": -24.895166397094727,
+      "logps/rejected": -98.50922393798828,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2309212684631348,
+      "rewards/margins": 8.308440208435059,
+      "rewards/rejected": -6.077519416809082,
+      "step": 1120
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9886363636363636e-07,
+      "logits/chosen": -1.9587482213974,
+      "logits/rejected": -1.5005452632904053,
+      "logps/chosen": -25.101999282836914,
+      "logps/rejected": -96.6336441040039,
+      "loss": 0.0011,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0761642456054688,
+      "rewards/margins": 8.064260482788086,
+      "rewards/rejected": -5.988096714019775,
+      "step": 1130
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9570707070707071e-07,
+      "logits/chosen": -2.1005218029022217,
+      "logits/rejected": -1.4738976955413818,
+      "logps/chosen": -26.019073486328125,
+      "logps/rejected": -101.70903015136719,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.189542293548584,
+      "rewards/margins": 8.359001159667969,
+      "rewards/rejected": -6.169457912445068,
+      "step": 1140
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9255050505050505e-07,
+      "logits/chosen": -1.9841840267181396,
+      "logits/rejected": -1.5473371744155884,
+      "logps/chosen": -25.14290428161621,
+      "logps/rejected": -98.56291198730469,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1460330486297607,
+      "rewards/margins": 8.121552467346191,
+      "rewards/rejected": -5.975518703460693,
+      "step": 1150
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.8939393939393938e-07,
+      "logits/chosen": -2.0372636318206787,
+      "logits/rejected": -1.4796743392944336,
+      "logps/chosen": -23.67437744140625,
+      "logps/rejected": -99.41020202636719,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.229948043823242,
+      "rewards/margins": 8.35695743560791,
+      "rewards/rejected": -6.12700891494751,
+      "step": 1160
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.862373737373737e-07,
+      "logits/chosen": -2.0319085121154785,
+      "logits/rejected": -1.5343225002288818,
+      "logps/chosen": -26.161046981811523,
+      "logps/rejected": -104.3691177368164,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0159668922424316,
+      "rewards/margins": 8.410650253295898,
+      "rewards/rejected": -6.394682884216309,
+      "step": 1170
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.830808080808081e-07,
+      "logits/chosen": -1.913827896118164,
+      "logits/rejected": -1.3936421871185303,
+      "logps/chosen": -23.682735443115234,
+      "logps/rejected": -99.6613998413086,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0931787490844727,
+      "rewards/margins": 8.25574779510498,
+      "rewards/rejected": -6.162568092346191,
+      "step": 1180
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.7992424242424243e-07,
+      "logits/chosen": -2.0480942726135254,
+      "logits/rejected": -1.497283697128296,
+      "logps/chosen": -27.119802474975586,
+      "logps/rejected": -96.8021469116211,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.116119861602783,
+      "rewards/margins": 8.150800704956055,
+      "rewards/rejected": -6.0346808433532715,
+      "step": 1190
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.7676767676767676e-07,
+      "logits/chosen": -2.047286033630371,
+      "logits/rejected": -1.53057861328125,
+      "logps/chosen": -24.611684799194336,
+      "logps/rejected": -101.38153076171875,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.130047559738159,
+      "rewards/margins": 8.328054428100586,
+      "rewards/rejected": -6.198008060455322,
+      "step": 1200
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.736111111111111e-07,
+      "logits/chosen": -2.0157628059387207,
+      "logits/rejected": -1.4612407684326172,
+      "logps/chosen": -26.134784698486328,
+      "logps/rejected": -96.5627212524414,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.109575033187866,
+      "rewards/margins": 8.016804695129395,
+      "rewards/rejected": -5.907229423522949,
+      "step": 1210
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.7045454545454543e-07,
+      "logits/chosen": -1.988457441329956,
+      "logits/rejected": -1.5492212772369385,
+      "logps/chosen": -25.654382705688477,
+      "logps/rejected": -99.71517181396484,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.066452741622925,
+      "rewards/margins": 8.255013465881348,
+      "rewards/rejected": -6.18856143951416,
+      "step": 1220
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.672979797979798e-07,
+      "logits/chosen": -1.9422584772109985,
+      "logits/rejected": -1.5462502241134644,
+      "logps/chosen": -26.08441162109375,
+      "logps/rejected": -98.68343353271484,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.143643617630005,
+      "rewards/margins": 8.192874908447266,
+      "rewards/rejected": -6.049231052398682,
+      "step": 1230
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.6414141414141414e-07,
+      "logits/chosen": -2.005826473236084,
+      "logits/rejected": -1.4449385404586792,
+      "logps/chosen": -23.385013580322266,
+      "logps/rejected": -100.96165466308594,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.224550724029541,
+      "rewards/margins": 8.504047393798828,
+      "rewards/rejected": -6.279496192932129,
+      "step": 1240
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.6098484848484847e-07,
+      "logits/chosen": -2.0005924701690674,
+      "logits/rejected": -1.5509334802627563,
+      "logps/chosen": -24.148326873779297,
+      "logps/rejected": -99.8790512084961,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1932854652404785,
+      "rewards/margins": 8.369304656982422,
+      "rewards/rejected": -6.176019668579102,
+      "step": 1250
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.578282828282828e-07,
+      "logits/chosen": -1.9986722469329834,
+      "logits/rejected": -1.4378080368041992,
+      "logps/chosen": -24.670684814453125,
+      "logps/rejected": -101.37323760986328,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.092477560043335,
+      "rewards/margins": 8.328160285949707,
+      "rewards/rejected": -6.235684394836426,
+      "step": 1260
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.5467171717171717e-07,
+      "logits/chosen": -1.9530527591705322,
+      "logits/rejected": -1.655870795249939,
+      "logps/chosen": -24.48198127746582,
+      "logps/rejected": -103.50440979003906,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1231515407562256,
+      "rewards/margins": 8.499826431274414,
+      "rewards/rejected": -6.376674652099609,
+      "step": 1270
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.5151515151515152e-07,
+      "logits/chosen": -1.9724531173706055,
+      "logits/rejected": -1.6001970767974854,
+      "logps/chosen": -25.00448226928711,
+      "logps/rejected": -96.61920928955078,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2021608352661133,
+      "rewards/margins": 8.14022159576416,
+      "rewards/rejected": -5.938061714172363,
+      "step": 1280
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4835858585858586e-07,
+      "logits/chosen": -2.0258517265319824,
+      "logits/rejected": -1.4299719333648682,
+      "logps/chosen": -25.8868465423584,
+      "logps/rejected": -104.29194641113281,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.160093307495117,
+      "rewards/margins": 8.549181938171387,
+      "rewards/rejected": -6.3890886306762695,
+      "step": 1290
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.452020202020202e-07,
+      "logits/chosen": -1.9465224742889404,
+      "logits/rejected": -1.456637978553772,
+      "logps/chosen": -23.794294357299805,
+      "logps/rejected": -100.16304016113281,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1781158447265625,
+      "rewards/margins": 8.355245590209961,
+      "rewards/rejected": -6.17712926864624,
+      "step": 1300
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4204545454545455e-07,
+      "logits/chosen": -1.9264423847198486,
+      "logits/rejected": -1.4020578861236572,
+      "logps/chosen": -27.03515625,
+      "logps/rejected": -99.86031341552734,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.104663372039795,
+      "rewards/margins": 8.347994804382324,
+      "rewards/rejected": -6.243332386016846,
+      "step": 1310
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.3888888888888888e-07,
+      "logits/chosen": -1.9269273281097412,
+      "logits/rejected": -1.378004550933838,
+      "logps/chosen": -25.283573150634766,
+      "logps/rejected": -97.65699768066406,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1022462844848633,
+      "rewards/margins": 8.27086067199707,
+      "rewards/rejected": -6.168614387512207,
+      "step": 1320
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3573232323232324e-07,
+      "logits/chosen": -1.9377596378326416,
+      "logits/rejected": -1.3735034465789795,
+      "logps/chosen": -24.710702896118164,
+      "logps/rejected": -105.88826751708984,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.202232837677002,
+      "rewards/margins": 8.641754150390625,
+      "rewards/rejected": -6.439520359039307,
+      "step": 1330
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.3257575757575757e-07,
+      "logits/chosen": -1.9057424068450928,
+      "logits/rejected": -1.3975918292999268,
+      "logps/chosen": -24.841745376586914,
+      "logps/rejected": -100.97312927246094,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1855947971343994,
+      "rewards/margins": 8.381650924682617,
+      "rewards/rejected": -6.196055889129639,
+      "step": 1340
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.294191919191919e-07,
+      "logits/chosen": -1.888952612876892,
+      "logits/rejected": -1.3854330778121948,
+      "logps/chosen": -27.225393295288086,
+      "logps/rejected": -96.71887969970703,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1763968467712402,
+      "rewards/margins": 8.234212875366211,
+      "rewards/rejected": -6.057814598083496,
+      "step": 1350
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.2626262626262626e-07,
+      "logits/chosen": -1.9558947086334229,
+      "logits/rejected": -1.4123071432113647,
+      "logps/chosen": -27.773752212524414,
+      "logps/rejected": -99.15750122070312,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.04634428024292,
+      "rewards/margins": 8.118974685668945,
+      "rewards/rejected": -6.072630405426025,
+      "step": 1360
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.231060606060606e-07,
+      "logits/chosen": -1.9197349548339844,
+      "logits/rejected": -1.4441031217575073,
+      "logps/chosen": -26.568771362304688,
+      "logps/rejected": -103.49000549316406,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0987703800201416,
+      "rewards/margins": 8.5416841506958,
+      "rewards/rejected": -6.442914009094238,
+      "step": 1370
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.1994949494949495e-07,
+      "logits/chosen": -1.9039863348007202,
+      "logits/rejected": -1.4260241985321045,
+      "logps/chosen": -27.353927612304688,
+      "logps/rejected": -104.0919418334961,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.091087579727173,
+      "rewards/margins": 8.414112091064453,
+      "rewards/rejected": -6.323023796081543,
+      "step": 1380
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.1679292929292928e-07,
+      "logits/chosen": -1.8462800979614258,
+      "logits/rejected": -1.478140115737915,
+      "logps/chosen": -26.750728607177734,
+      "logps/rejected": -102.26488494873047,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.132601022720337,
+      "rewards/margins": 8.408994674682617,
+      "rewards/rejected": -6.276393890380859,
+      "step": 1390
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.1363636363636363e-07,
+      "logits/chosen": -1.8644853830337524,
+      "logits/rejected": -1.5662178993225098,
+      "logps/chosen": -23.476882934570312,
+      "logps/rejected": -101.57548522949219,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1439037322998047,
+      "rewards/margins": 8.44482707977295,
+      "rewards/rejected": -6.3009233474731445,
+      "step": 1400
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.1047979797979798e-07,
+      "logits/chosen": -1.8527309894561768,
+      "logits/rejected": -1.4847913980484009,
+      "logps/chosen": -25.65964126586914,
+      "logps/rejected": -102.22838592529297,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.148160457611084,
+      "rewards/margins": 8.409781455993652,
+      "rewards/rejected": -6.261620998382568,
+      "step": 1410
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.0732323232323232e-07,
+      "logits/chosen": -1.9628047943115234,
+      "logits/rejected": -1.3477706909179688,
+      "logps/chosen": -24.940982818603516,
+      "logps/rejected": -101.76347351074219,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1268465518951416,
+      "rewards/margins": 8.330877304077148,
+      "rewards/rejected": -6.2040300369262695,
+      "step": 1420
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.0416666666666667e-07,
+      "logits/chosen": -1.9355055093765259,
+      "logits/rejected": -1.4377237558364868,
+      "logps/chosen": -25.779720306396484,
+      "logps/rejected": -101.36628723144531,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2160532474517822,
+      "rewards/margins": 8.526556968688965,
+      "rewards/rejected": -6.31050443649292,
+      "step": 1430
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.01010101010101e-07,
+      "logits/chosen": -1.9768778085708618,
+      "logits/rejected": -1.3711049556732178,
+      "logps/chosen": -26.17185401916504,
+      "logps/rejected": -103.18241119384766,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.193679094314575,
+      "rewards/margins": 8.478619575500488,
+      "rewards/rejected": -6.284939765930176,
+      "step": 1440
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 9.785353535353536e-08,
+      "logits/chosen": -1.872606635093689,
+      "logits/rejected": -1.3800017833709717,
+      "logps/chosen": -27.725723266601562,
+      "logps/rejected": -105.1909408569336,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2797818183898926,
+      "rewards/margins": 8.561399459838867,
+      "rewards/rejected": -6.281617641448975,
+      "step": 1450
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 9.469696969696969e-08,
+      "logits/chosen": -1.921291708946228,
+      "logits/rejected": -1.3621512651443481,
+      "logps/chosen": -26.722686767578125,
+      "logps/rejected": -100.7763442993164,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1938297748565674,
+      "rewards/margins": 8.335715293884277,
+      "rewards/rejected": -6.141885280609131,
+      "step": 1460
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 9.154040404040405e-08,
+      "logits/chosen": -1.8397643566131592,
+      "logits/rejected": -1.449937343597412,
+      "logps/chosen": -24.311527252197266,
+      "logps/rejected": -103.50492858886719,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.17022705078125,
+      "rewards/margins": 8.600130081176758,
+      "rewards/rejected": -6.42990255355835,
+      "step": 1470
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.838383838383838e-08,
+      "logits/chosen": -1.8616399765014648,
+      "logits/rejected": -1.4163103103637695,
+      "logps/chosen": -24.417814254760742,
+      "logps/rejected": -100.59847259521484,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1278107166290283,
+      "rewards/margins": 8.353968620300293,
+      "rewards/rejected": -6.22615909576416,
+      "step": 1480
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.522727272727271e-08,
+      "logits/chosen": -1.8747402429580688,
+      "logits/rejected": -1.3399763107299805,
+      "logps/chosen": -24.992324829101562,
+      "logps/rejected": -100.95648956298828,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1101977825164795,
+      "rewards/margins": 8.302995681762695,
+      "rewards/rejected": -6.192798614501953,
+      "step": 1490
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.207070707070707e-08,
+      "logits/chosen": -1.9132680892944336,
+      "logits/rejected": -1.4688284397125244,
+      "logps/chosen": -22.073471069335938,
+      "logps/rejected": -100.51573181152344,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.18825101852417,
+      "rewards/margins": 8.584248542785645,
+      "rewards/rejected": -6.395996570587158,
+      "step": 1500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 7.89141414141414e-08,
+      "logits/chosen": -1.9575687646865845,
+      "logits/rejected": -1.3903148174285889,
+      "logps/chosen": -27.41025161743164,
+      "logps/rejected": -103.32008361816406,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1470372676849365,
+      "rewards/margins": 8.490819931030273,
+      "rewards/rejected": -6.3437819480896,
+      "step": 1510
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 7.575757575757576e-08,
+      "logits/chosen": -1.9653770923614502,
+      "logits/rejected": -1.489027500152588,
+      "logps/chosen": -25.78949546813965,
+      "logps/rejected": -101.89344787597656,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1554603576660156,
+      "rewards/margins": 8.434431076049805,
+      "rewards/rejected": -6.278970241546631,
+      "step": 1520
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 7.26010101010101e-08,
+      "logits/chosen": -1.756226897239685,
+      "logits/rejected": -1.4216405153274536,
+      "logps/chosen": -29.357797622680664,
+      "logps/rejected": -99.86305236816406,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.11063551902771,
+      "rewards/margins": 8.357722282409668,
+      "rewards/rejected": -6.247086524963379,
+      "step": 1530
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.944444444444444e-08,
+      "logits/chosen": -1.8492367267608643,
+      "logits/rejected": -1.3363265991210938,
+      "logps/chosen": -25.58880043029785,
+      "logps/rejected": -103.36441802978516,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.101243019104004,
+      "rewards/margins": 8.492372512817383,
+      "rewards/rejected": -6.391129493713379,
+      "step": 1540
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.628787878787879e-08,
+      "logits/chosen": -1.8674955368041992,
+      "logits/rejected": -1.373772382736206,
+      "logps/chosen": -26.437780380249023,
+      "logps/rejected": -105.77347564697266,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2881102561950684,
+      "rewards/margins": 8.857427597045898,
+      "rewards/rejected": -6.569317817687988,
+      "step": 1550
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.313131313131313e-08,
+      "logits/chosen": -1.7903921604156494,
+      "logits/rejected": -1.2768902778625488,
+      "logps/chosen": -24.630321502685547,
+      "logps/rejected": -100.84373474121094,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.17745304107666,
+      "rewards/margins": 8.567586898803711,
+      "rewards/rejected": -6.390133857727051,
+      "step": 1560
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.997474747474748e-08,
+      "logits/chosen": -1.8087184429168701,
+      "logits/rejected": -1.409135103225708,
+      "logps/chosen": -25.476566314697266,
+      "logps/rejected": -102.30812072753906,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.134692430496216,
+      "rewards/margins": 8.463730812072754,
+      "rewards/rejected": -6.329038143157959,
+      "step": 1570
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.6818181818181815e-08,
+      "logits/chosen": -1.7994487285614014,
+      "logits/rejected": -1.3598864078521729,
+      "logps/chosen": -25.109106063842773,
+      "logps/rejected": -103.5053482055664,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1917526721954346,
+      "rewards/margins": 8.73997688293457,
+      "rewards/rejected": -6.54822301864624,
+      "step": 1580
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.366161616161616e-08,
+      "logits/chosen": -1.7967250347137451,
+      "logits/rejected": -1.2914398908615112,
+      "logps/chosen": -23.77326202392578,
+      "logps/rejected": -101.91696166992188,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2252016067504883,
+      "rewards/margins": 8.633685111999512,
+      "rewards/rejected": -6.408483982086182,
+      "step": 1590
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.05050505050505e-08,
+      "logits/chosen": -1.7815173864364624,
+      "logits/rejected": -1.3278355598449707,
+      "logps/chosen": -26.007434844970703,
+      "logps/rejected": -102.10801696777344,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1467533111572266,
+      "rewards/margins": 8.515238761901855,
+      "rewards/rejected": -6.368485450744629,
+      "step": 1600
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.7348484848484845e-08,
+      "logits/chosen": -1.8696527481079102,
+      "logits/rejected": -1.1647627353668213,
+      "logps/chosen": -26.391366958618164,
+      "logps/rejected": -102.64463806152344,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1689181327819824,
+      "rewards/margins": 8.574686050415039,
+      "rewards/rejected": -6.40576696395874,
+      "step": 1610
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.419191919191919e-08,
+      "logits/chosen": -1.8303635120391846,
+      "logits/rejected": -1.4279205799102783,
+      "logps/chosen": -25.891845703125,
+      "logps/rejected": -102.86759948730469,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1555731296539307,
+      "rewards/margins": 8.519299507141113,
+      "rewards/rejected": -6.3637261390686035,
+      "step": 1620
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.1035353535353536e-08,
+      "logits/chosen": -1.8141838312149048,
+      "logits/rejected": -1.3252172470092773,
+      "logps/chosen": -25.70889663696289,
+      "logps/rejected": -101.9534683227539,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1951491832733154,
+      "rewards/margins": 8.575303077697754,
+      "rewards/rejected": -6.380153656005859,
+      "step": 1630
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.787878787878788e-08,
+      "logits/chosen": -1.7714385986328125,
+      "logits/rejected": -1.390956997871399,
+      "logps/chosen": -27.153860092163086,
+      "logps/rejected": -102.859130859375,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.125636577606201,
+      "rewards/margins": 8.567439079284668,
+      "rewards/rejected": -6.441802024841309,
+      "step": 1640
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.472222222222222e-08,
+      "logits/chosen": -1.9194447994232178,
+      "logits/rejected": -1.4019012451171875,
+      "logps/chosen": -24.547632217407227,
+      "logps/rejected": -101.7523422241211,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2368247509002686,
+      "rewards/margins": 8.566001892089844,
+      "rewards/rejected": -6.329176902770996,
+      "step": 1650
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.1565656565656565e-08,
+      "logits/chosen": -1.7740247249603271,
+      "logits/rejected": -1.1793501377105713,
+      "logps/chosen": -22.256084442138672,
+      "logps/rejected": -99.48062133789062,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2038888931274414,
+      "rewards/margins": 8.503969192504883,
+      "rewards/rejected": -6.3000807762146,
+      "step": 1660
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.8409090909090908e-08,
+      "logits/chosen": -1.8418700695037842,
+      "logits/rejected": -1.3521521091461182,
+      "logps/chosen": -25.66836929321289,
+      "logps/rejected": -104.65287780761719,
+      "loss": 0.0007,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.198652744293213,
+      "rewards/margins": 8.75655460357666,
+      "rewards/rejected": -6.557901859283447,
+      "step": 1670
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 2.525252525252525e-08,
+      "logits/chosen": -1.8341500759124756,
+      "logits/rejected": -1.3680517673492432,
+      "logps/chosen": -24.97684097290039,
+      "logps/rejected": -105.42335510253906,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2152397632598877,
+      "rewards/margins": 8.671110153198242,
+      "rewards/rejected": -6.45587158203125,
+      "step": 1680
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.2095959595959595e-08,
+      "logits/chosen": -1.7939857244491577,
+      "logits/rejected": -1.265699028968811,
+      "logps/chosen": -27.59762191772461,
+      "logps/rejected": -108.02186584472656,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.129595994949341,
+      "rewards/margins": 8.90428352355957,
+      "rewards/rejected": -6.77468729019165,
+      "step": 1690
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.893939393939394e-08,
+      "logits/chosen": -1.8354851007461548,
+      "logits/rejected": -1.2298798561096191,
+      "logps/chosen": -26.194656372070312,
+      "logps/rejected": -105.52481842041016,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0637497901916504,
+      "rewards/margins": 8.558916091918945,
+      "rewards/rejected": -6.4951653480529785,
+      "step": 1700
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.5782828282828283e-08,
+      "logits/chosen": -1.7558263540267944,
+      "logits/rejected": -1.2703979015350342,
+      "logps/chosen": -25.03870391845703,
+      "logps/rejected": -102.92311096191406,
+      "loss": 0.0006,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2000644207000732,
+      "rewards/margins": 8.537808418273926,
+      "rewards/rejected": -6.337743282318115,
+      "step": 1710
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2626262626262625e-08,
+      "logits/chosen": -1.8095840215682983,
+      "logits/rejected": -1.2369449138641357,
+      "logps/chosen": -24.758800506591797,
+      "logps/rejected": -104.48983001708984,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2792153358459473,
+      "rewards/margins": 8.828054428100586,
+      "rewards/rejected": -6.5488386154174805,
+      "step": 1720
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.46969696969697e-09,
+      "logits/chosen": -1.80984365940094,
+      "logits/rejected": -1.1332688331604004,
+      "logps/chosen": -27.860952377319336,
+      "logps/rejected": -102.5757827758789,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.0335052013397217,
+      "rewards/margins": 8.37595272064209,
+      "rewards/rejected": -6.342446327209473,
+      "step": 1730
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.313131313131312e-09,
+      "logits/chosen": -1.834185004234314,
+      "logits/rejected": -1.135471224784851,
+      "logps/chosen": -26.846776962280273,
+      "logps/rejected": -105.81257629394531,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.1772818565368652,
+      "rewards/margins": 8.714735984802246,
+      "rewards/rejected": -6.537454128265381,
+      "step": 1740
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.156565656565656e-09,
+      "logits/chosen": -1.8649733066558838,
+      "logits/rejected": -1.2481504678726196,
+      "logps/chosen": -25.80572509765625,
+      "logps/rejected": -105.42039489746094,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2440056800842285,
+      "rewards/margins": 8.69550609588623,
+      "rewards/rejected": -6.45150089263916,
+      "step": 1750
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0,
+      "logits/chosen": -1.7620856761932373,
+      "logits/rejected": -1.2375068664550781,
+      "logps/chosen": -24.10125732421875,
+      "logps/rejected": -105.0677490234375,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 2.2944893836975098,
+      "rewards/margins": 8.900548934936523,
+      "rewards/rejected": -6.606060028076172,
+      "step": 1760
+    },
+    {
+      "epoch": 1.0,
+      "eval_logits/chosen": -1.7432786226272583,
+      "eval_logits/rejected": -1.3963388204574585,
+      "eval_logps/chosen": -26.21647834777832,
+      "eval_logps/rejected": -101.43244934082031,
+      "eval_loss": 0.00026922515826299787,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": 1.9645460844039917,
+      "eval_rewards/margins": 8.17833423614502,
+      "eval_rewards/rejected": -6.2137885093688965,
+      "eval_runtime": 5.059,
+      "eval_samples_per_second": 19.767,
+      "eval_steps_per_second": 1.384,
+      "step": 1760
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1760,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "total_flos": 0.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

mistral_7b_step_dpo_var2/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d79e833d94f6ce9492d3674f96bf182b07ad4fa0fe024e90faca100dd863b830
+size 5880

mistral_7b_step_dpo_var2/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)