Upload folder using huggingface_hub

added_tokens.json

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

config.json

@@ -0,0 +1,29 @@
+{
+ "_name_or_path": "Qwen/Qwen2.5-7B",
+ "architectures": [
+ "Qwen2ForCausalLM"
+ ],
+ "attention_dropout": 0.0,
+ "eos_token_id": 151643,
+ "hidden_act": "silu",
+ "hidden_size": 3584,
+ "initializer_range": 0.02,
+ "intermediate_size": 18944,
+ "max_position_embeddings": 131072,
+ "max_window_layers": 28,
+ "model_type": "qwen2",
+ "num_attention_heads": 28,
+ "num_hidden_layers": 28,
+ "num_key_value_heads": 4,
+ "rms_norm_eps": 1e-06,
+ "rope_scaling": null,
+ "rope_theta": 1000000.0,
+ "sliding_window": null,
+ "tie_word_embeddings": false,
+ "torch_dtype": "bfloat16",
+ "transformers_version": "4.45.1",
+ "use_cache": false,
+ "use_mrope": false,
+ "use_sliding_window": false,
+ "vocab_size": 152064
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+ "bos_token_id": 151643,
+ "do_sample": true,
+ "eos_token_id": 151643,
+ "max_new_tokens": 2048,
+ "transformers_version": "4.45.1"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c76fa96a7451ac044ab0bbfec547783bde3a1aca86cb1357d390149c29ea2527
+size 4877660776

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3991f71374e83952e55049260e2933a6c78c337e78790ac798528b5a9d49834
+size 4932751008

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9cfb2a1a963db89a6841e6d3415b766076e4f59ae068d788fcdb3bcd713ecc4
+size 4330865200

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b8aa26a04efea407b900b2ccd8c14eaaa71adb68cbf77a28fe6678e39014a68d
+size 1089994880

model.safetensors.index.json

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

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1,207 @@
+{
+ "add_bos_token": false,
+ "add_prefix_space": false,
+ "added_tokens_decoder": {
+ "151643": {
+ "content": "<|endoftext|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151644": {
+ "content": "<|im_start|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151645": {
+ "content": "<|im_end|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151646": {
+ "content": "<|object_ref_start|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151647": {
+ "content": "<|object_ref_end|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151648": {
+ "content": "<|box_start|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151649": {
+ "content": "<|box_end|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151650": {
+ "content": "<|quad_start|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151651": {
+ "content": "<|quad_end|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151652": {
+ "content": "<|vision_start|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151653": {
+ "content": "<|vision_end|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151654": {
+ "content": "<|vision_pad|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151655": {
+ "content": "<|image_pad|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151656": {
+ "content": "<|video_pad|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": true
+ },
+ "151657": {
+ "content": "<tool_call>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151658": {
+ "content": "</tool_call>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151659": {
+ "content": "<|fim_prefix|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151660": {
+ "content": "<|fim_middle|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151661": {
+ "content": "<|fim_suffix|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151662": {
+ "content": "<|fim_pad|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151663": {
+ "content": "<|repo_name|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ },
+ "151664": {
+ "content": "<|file_sep|>",
+ "lstrip": false,
+ "normalized": false,
+ "rstrip": false,
+ "single_word": false,
+ "special": false
+ }
+ },
+ "additional_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "bos_token": null,
+ "chat_template": "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}

trainer_state.json

@@ -0,0 +1,2706 @@
+{
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.2539152223846315,
+ "eval_steps": 75,
+ "global_step": 375,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "epoch": 0.003341687552213868,
+ "grad_norm": 9.422656730262974,
+ "learning_rate": 1.6666666666666667e-06,
+ "loss": 1.4155,
+ "step": 1
+ },
+ {
+ "epoch": 0.003341687552213868,
+ "eval_loss": 1.8004764318466187,
+ "eval_runtime": 49.231,
+ "eval_samples_per_second": 0.447,
+ "eval_steps_per_second": 0.122,
+ "step": 1
+ },
+ {
+ "epoch": 0.006683375104427736,
+ "grad_norm": 7.9592686195064655,
+ "learning_rate": 3.3333333333333333e-06,
+ "loss": 1.4836,
+ "step": 2
+ },
+ {
+ "epoch": 0.010025062656641603,
+ "grad_norm": 11.08227927061808,
+ "learning_rate": 5e-06,
+ "loss": 1.568,
+ "step": 3
+ },
+ {
+ "epoch": 0.013366750208855471,
+ "grad_norm": 9.054404458978134,
+ "learning_rate": 6.666666666666667e-06,
+ "loss": 1.4902,
+ "step": 4
+ },
+ {
+ "epoch": 0.01670843776106934,
+ "grad_norm": 4.814739146640832,
+ "learning_rate": 8.333333333333334e-06,
+ "loss": 1.4581,
+ "step": 5
+ },
+ {
+ "epoch": 0.020050125313283207,
+ "grad_norm": 3.370760758247303,
+ "learning_rate": 1e-05,
+ "loss": 1.4943,
+ "step": 6
+ },
+ {
+ "epoch": 0.023391812865497075,
+ "grad_norm": 4.15650648864708,
+ "learning_rate": 1.1666666666666668e-05,
+ "loss": 1.3664,
+ "step": 7
+ },
+ {
+ "epoch": 0.026733500417710943,
+ "grad_norm": 4.168972921570349,
+ "learning_rate": 1.3333333333333333e-05,
+ "loss": 1.369,
+ "step": 8
+ },
+ {
+ "epoch": 0.03007518796992481,
+ "grad_norm": 2.5626426979817754,
+ "learning_rate": 1.5e-05,
+ "loss": 1.4485,
+ "step": 9
+ },
+ {
+ "epoch": 0.03341687552213868,
+ "grad_norm": 1266.5995492196223,
+ "learning_rate": 1.6666666666666667e-05,
+ "loss": 1.3401,
+ "step": 10
+ },
+ {
+ "epoch": 0.036758563074352546,
+ "grad_norm": 2.544199092180538,
+ "learning_rate": 1.8333333333333333e-05,
+ "loss": 1.3449,
+ "step": 11
+ },
+ {
+ "epoch": 0.040100250626566414,
+ "grad_norm": 2.2046489399946028,
+ "learning_rate": 2e-05,
+ "loss": 1.3165,
+ "step": 12
+ },
+ {
+ "epoch": 0.04344193817878028,
+ "grad_norm": 1.8293693281340389,
+ "learning_rate": 2.1666666666666667e-05,
+ "loss": 1.397,
+ "step": 13
+ },
+ {
+ "epoch": 0.04678362573099415,
+ "grad_norm": 2.149761301718473,
+ "learning_rate": 2.3333333333333336e-05,
+ "loss": 1.436,
+ "step": 14
+ },
+ {
+ "epoch": 0.05012531328320802,
+ "grad_norm": 2.1865938834852523,
+ "learning_rate": 2.5e-05,
+ "loss": 1.3629,
+ "step": 15
+ },
+ {
+ "epoch": 0.053467000835421885,
+ "grad_norm": 1.6718675362366882,
+ "learning_rate": 2.6666666666666667e-05,
+ "loss": 1.2787,
+ "step": 16
+ },
+ {
+ "epoch": 0.05680868838763575,
+ "grad_norm": 1.7469631110837789,
+ "learning_rate": 2.8333333333333335e-05,
+ "loss": 1.3666,
+ "step": 17
+ },
+ {
+ "epoch": 0.06015037593984962,
+ "grad_norm": 2.026182288959042,
+ "learning_rate": 3e-05,
+ "loss": 1.3905,
+ "step": 18
+ },
+ {
+ "epoch": 0.06349206349206349,
+ "grad_norm": 1.5601215827368777,
+ "learning_rate": 3.1666666666666666e-05,
+ "loss": 1.2615,
+ "step": 19
+ },
+ {
+ "epoch": 0.06683375104427736,
+ "grad_norm": 1.6239108972686256,
+ "learning_rate": 3.3333333333333335e-05,
+ "loss": 1.2865,
+ "step": 20
+ },
+ {
+ "epoch": 0.07017543859649122,
+ "grad_norm": 1.8043652555869882,
+ "learning_rate": 3.5e-05,
+ "loss": 1.3364,
+ "step": 21
+ },
+ {
+ "epoch": 0.07351712614870509,
+ "grad_norm": 1.418878317939115,
+ "learning_rate": 3.6666666666666666e-05,
+ "loss": 1.239,
+ "step": 22
+ },
+ {
+ "epoch": 0.07685881370091896,
+ "grad_norm": 1.641006294199728,
+ "learning_rate": 3.8333333333333334e-05,
+ "loss": 1.2206,
+ "step": 23
+ },
+ {
+ "epoch": 0.08020050125313283,
+ "grad_norm": 1.4565002852740556,
+ "learning_rate": 4e-05,
+ "loss": 1.3179,
+ "step": 24
+ },
+ {
+ "epoch": 0.0835421888053467,
+ "grad_norm": 1.2586337015066662,
+ "learning_rate": 4.166666666666667e-05,
+ "loss": 1.2097,
+ "step": 25
+ },
+ {
+ "epoch": 0.08688387635756056,
+ "grad_norm": 1.2904448628312173,
+ "learning_rate": 4.3333333333333334e-05,
+ "loss": 1.2804,
+ "step": 26
+ },
+ {
+ "epoch": 0.09022556390977443,
+ "grad_norm": 1.336149520030802,
+ "learning_rate": 4.5e-05,
+ "loss": 1.2905,
+ "step": 27
+ },
+ {
+ "epoch": 0.0935672514619883,
+ "grad_norm": 1.3805663163469717,
+ "learning_rate": 4.666666666666667e-05,
+ "loss": 1.2018,
+ "step": 28
+ },
+ {
+ "epoch": 0.09690893901420217,
+ "grad_norm": 1.3223829196481407,
+ "learning_rate": 4.8333333333333334e-05,
+ "loss": 1.3514,
+ "step": 29
+ },
+ {
+ "epoch": 0.10025062656641603,
+ "grad_norm": 1.117557068600307,
+ "learning_rate": 5e-05,
+ "loss": 1.3185,
+ "step": 30
+ },
+ {
+ "epoch": 0.1035923141186299,
+ "grad_norm": 1.3628329620488002,
+ "learning_rate": 4.999983587637969e-05,
+ "loss": 1.3184,
+ "step": 31
+ },
+ {
+ "epoch": 0.10693400167084377,
+ "grad_norm": 1.6900882409593405,
+ "learning_rate": 4.999934350767369e-05,
+ "loss": 1.2478,
+ "step": 32
+ },
+ {
+ "epoch": 0.11027568922305764,
+ "grad_norm": 1.1018164563546424,
+ "learning_rate": 4.999852290034673e-05,
+ "loss": 1.1818,
+ "step": 33
+ },
+ {
+ "epoch": 0.1136173767752715,
+ "grad_norm": 1.517905740455229,
+ "learning_rate": 4.9997374065173316e-05,
+ "loss": 1.241,
+ "step": 34
+ },
+ {
+ "epoch": 0.11695906432748537,
+ "grad_norm": 1.3462450662992849,
+ "learning_rate": 4.999589701723751e-05,
+ "loss": 1.4392,
+ "step": 35
+ },
+ {
+ "epoch": 0.12030075187969924,
+ "grad_norm": 1.332649775646042,
+ "learning_rate": 4.99940917759328e-05,
+ "loss": 1.3001,
+ "step": 36
+ },
+ {
+ "epoch": 0.12364243943191311,
+ "grad_norm": 1.3291037070625635,
+ "learning_rate": 4.9991958364961794e-05,
+ "loss": 1.2308,
+ "step": 37
+ },
+ {
+ "epoch": 0.12698412698412698,
+ "grad_norm": 1.42150501947453,
+ "learning_rate": 4.998949681233595e-05,
+ "loss": 1.2918,
+ "step": 38
+ },
+ {
+ "epoch": 0.13032581453634084,
+ "grad_norm": 1.074142668859534,
+ "learning_rate": 4.998670715037518e-05,
+ "loss": 1.278,
+ "step": 39
+ },
+ {
+ "epoch": 0.1336675020885547,
+ "grad_norm": 1.2920579933597587,
+ "learning_rate": 4.998358941570744e-05,
+ "loss": 1.3275,
+ "step": 40
+ },
+ {
+ "epoch": 0.13700918964076858,
+ "grad_norm": 1.237403262913235,
+ "learning_rate": 4.998014364926824e-05,
+ "loss": 1.2,
+ "step": 41
+ },
+ {
+ "epoch": 0.14035087719298245,
+ "grad_norm": 1.268092315378372,
+ "learning_rate": 4.997636989630011e-05,
+ "loss": 1.1777,
+ "step": 42
+ },
+ {
+ "epoch": 0.14369256474519632,
+ "grad_norm": 1.0207096180154704,
+ "learning_rate": 4.997226820635203e-05,
+ "loss": 1.2758,
+ "step": 43
+ },
+ {
+ "epoch": 0.14703425229741018,
+ "grad_norm": 1.3430934015213214,
+ "learning_rate": 4.99678386332787e-05,
+ "loss": 1.1911,
+ "step": 44
+ },
+ {
+ "epoch": 0.15037593984962405,
+ "grad_norm": 1.1541607076308344,
+ "learning_rate": 4.996308123523995e-05,
+ "loss": 1.2296,
+ "step": 45
+ },
+ {
+ "epoch": 0.15371762740183792,
+ "grad_norm": 1.2443113266651544,
+ "learning_rate": 4.9957996074699884e-05,
+ "loss": 1.2381,
+ "step": 46
+ },
+ {
+ "epoch": 0.1570593149540518,
+ "grad_norm": 1.0652313229665111,
+ "learning_rate": 4.995258321842611e-05,
+ "loss": 1.1833,
+ "step": 47
+ },
+ {
+ "epoch": 0.16040100250626566,
+ "grad_norm": 1.1543672819429747,
+ "learning_rate": 4.9946842737488815e-05,
+ "loss": 1.2179,
+ "step": 48
+ },
+ {
+ "epoch": 0.16374269005847952,
+ "grad_norm": 1.1021407909957812,
+ "learning_rate": 4.994077470725989e-05,
+ "loss": 1.2463,
+ "step": 49
+ },
+ {
+ "epoch": 0.1670843776106934,
+ "grad_norm": 1.1322090567817211,
+ "learning_rate": 4.993437920741191e-05,
+ "loss": 1.1531,
+ "step": 50
+ },
+ {
+ "epoch": 0.17042606516290726,
+ "grad_norm": 1.272598115993271,
+ "learning_rate": 4.992765632191706e-05,
+ "loss": 1.2044,
+ "step": 51
+ },
+ {
+ "epoch": 0.17376775271512113,
+ "grad_norm": 1.2820610442976028,
+ "learning_rate": 4.992060613904611e-05,
+ "loss": 1.2833,
+ "step": 52
+ },
+ {
+ "epoch": 0.177109440267335,
+ "grad_norm": 1.03946480935167,
+ "learning_rate": 4.991322875136716e-05,
+ "loss": 1.2043,
+ "step": 53
+ },
+ {
+ "epoch": 0.18045112781954886,
+ "grad_norm": 1.4100014931322022,
+ "learning_rate": 4.990552425574452e-05,
+ "loss": 1.2572,
+ "step": 54
+ },
+ {
+ "epoch": 0.18379281537176273,
+ "grad_norm": 1.1044198987392109,
+ "learning_rate": 4.989749275333735e-05,
+ "loss": 1.173,
+ "step": 55
+ },
+ {
+ "epoch": 0.1871345029239766,
+ "grad_norm": 1.2029957139213148,
+ "learning_rate": 4.9889134349598384e-05,
+ "loss": 1.2086,
+ "step": 56
+ },
+ {
+ "epoch": 0.19047619047619047,
+ "grad_norm": 1.1209588633953635,
+ "learning_rate": 4.988044915427257e-05,
+ "loss": 1.2347,
+ "step": 57
+ },
+ {
+ "epoch": 0.19381787802840433,
+ "grad_norm": 1.1050690405175774,
+ "learning_rate": 4.9871437281395525e-05,
+ "loss": 1.2299,
+ "step": 58
+ },
+ {
+ "epoch": 0.1971595655806182,
+ "grad_norm": 1.121471860445537,
+ "learning_rate": 4.986209884929218e-05,
+ "loss": 1.1989,
+ "step": 59
+ },
+ {
+ "epoch": 0.20050125313283207,
+ "grad_norm": 1.1616230448507037,
+ "learning_rate": 4.98524339805751e-05,
+ "loss": 1.392,
+ "step": 60
+ },
+ {
+ "epoch": 0.20384294068504594,
+ "grad_norm": 1.091249962757003,
+ "learning_rate": 4.984244280214294e-05,
+ "loss": 1.2663,
+ "step": 61
+ },
+ {
+ "epoch": 0.2071846282372598,
+ "grad_norm": 1.1157809777653205,
+ "learning_rate": 4.983212544517878e-05,
+ "loss": 1.2751,
+ "step": 62
+ },
+ {
+ "epoch": 0.21052631578947367,
+ "grad_norm": 1.2254356056906608,
+ "learning_rate": 4.9821482045148375e-05,
+ "loss": 1.2468,
+ "step": 63
+ },
+ {
+ "epoch": 0.21386800334168754,
+ "grad_norm": 1.0405533895078984,
+ "learning_rate": 4.98105127417984e-05,
+ "loss": 1.1853,
+ "step": 64
+ },
+ {
+ "epoch": 0.2172096908939014,
+ "grad_norm": 1.1342961308693198,
+ "learning_rate": 4.9799217679154584e-05,
+ "loss": 1.2196,
+ "step": 65
+ },
+ {
+ "epoch": 0.22055137844611528,
+ "grad_norm": 1.1946061028510457,
+ "learning_rate": 4.978759700551986e-05,
+ "loss": 1.1807,
+ "step": 66
+ },
+ {
+ "epoch": 0.22389306599832914,
+ "grad_norm": 1.1751408913504486,
+ "learning_rate": 4.977565087347239e-05,
+ "loss": 1.2289,
+ "step": 67
+ },
+ {
+ "epoch": 0.227234753550543,
+ "grad_norm": 1.3191541216438447,
+ "learning_rate": 4.976337943986357e-05,
+ "loss": 1.214,
+ "step": 68
+ },
+ {
+ "epoch": 0.23057644110275688,
+ "grad_norm": 1.1444973785656187,
+ "learning_rate": 4.975078286581596e-05,
+ "loss": 1.2287,
+ "step": 69
+ },
+ {
+ "epoch": 0.23391812865497075,
+ "grad_norm": 1.2184149728817883,
+ "learning_rate": 4.97378613167212e-05,
+ "loss": 1.3791,
+ "step": 70
+ },
+ {
+ "epoch": 0.23725981620718462,
+ "grad_norm": 1.1647377307572258,
+ "learning_rate": 4.9724614962237803e-05,
+ "loss": 1.1693,
+ "step": 71
+ },
+ {
+ "epoch": 0.24060150375939848,
+ "grad_norm": 1.1110859065431626,
+ "learning_rate": 4.971104397628893e-05,
+ "loss": 1.2154,
+ "step": 72
+ },
+ {
+ "epoch": 0.24394319131161235,
+ "grad_norm": 1.2567000920579476,
+ "learning_rate": 4.9697148537060136e-05,
+ "loss": 1.1611,
+ "step": 73
+ },
+ {
+ "epoch": 0.24728487886382622,
+ "grad_norm": 1.1550950387887173,
+ "learning_rate": 4.9682928826997e-05,
+ "loss": 1.257,
+ "step": 74
+ },
+ {
+ "epoch": 0.2506265664160401,
+ "grad_norm": 1.0281075297368318,
+ "learning_rate": 4.966838503280275e-05,
+ "loss": 1.1868,
+ "step": 75
+ },
+ {
+ "epoch": 0.2506265664160401,
+ "eval_loss": 1.4157997369766235,
+ "eval_runtime": 48.5608,
+ "eval_samples_per_second": 0.453,
+ "eval_steps_per_second": 0.124,
+ "step": 75
+ },
+ {
+ "epoch": 0.25396825396825395,
+ "grad_norm": 1.1032178941225772,
+ "learning_rate": 4.96535173454358e-05,
+ "loss": 1.1861,
+ "step": 76
+ },
+ {
+ "epoch": 0.2573099415204678,
+ "grad_norm": 1.093369250778169,
+ "learning_rate": 4.9638325960107235e-05,
+ "loss": 1.2555,
+ "step": 77
+ },
+ {
+ "epoch": 0.2606516290726817,
+ "grad_norm": 1.1413869539274288,
+ "learning_rate": 4.962281107627828e-05,
+ "loss": 1.2892,
+ "step": 78
+ },
+ {
+ "epoch": 0.26399331662489556,
+ "grad_norm": 1.0532164721638475,
+ "learning_rate": 4.960697289765764e-05,
+ "loss": 1.2131,
+ "step": 79
+ },
+ {
+ "epoch": 0.2673350041771094,
+ "grad_norm": 1.0579041886059877,
+ "learning_rate": 4.959081163219885e-05,
+ "loss": 1.2693,
+ "step": 80
+ },
+ {
+ "epoch": 0.2706766917293233,
+ "grad_norm": 1.1312038750809115,
+ "learning_rate": 4.957432749209755e-05,
+ "loss": 1.2504,
+ "step": 81
+ },
+ {
+ "epoch": 0.27401837928153716,
+ "grad_norm": 1.0252412589334967,
+ "learning_rate": 4.955752069378867e-05,
+ "loss": 1.1226,
+ "step": 82
+ },
+ {
+ "epoch": 0.27736006683375103,
+ "grad_norm": 1.0668623234540449,
+ "learning_rate": 4.9540391457943626e-05,
+ "loss": 1.2471,
+ "step": 83
+ },
+ {
+ "epoch": 0.2807017543859649,
+ "grad_norm": 1.1247331389355435,
+ "learning_rate": 4.9522940009467385e-05,
+ "loss": 1.3092,
+ "step": 84
+ },
+ {
+ "epoch": 0.28404344193817876,
+ "grad_norm": 1.0404338917876472,
+ "learning_rate": 4.9505166577495546e-05,
+ "loss": 1.2724,
+ "step": 85
+ },
+ {
+ "epoch": 0.28738512949039263,
+ "grad_norm": 1.0666603820625944,
+ "learning_rate": 4.9487071395391304e-05,
+ "loss": 1.212,
+ "step": 86
+ },
+ {
+ "epoch": 0.2907268170426065,
+ "grad_norm": 1.028172195805746,
+ "learning_rate": 4.946865470074241e-05,
+ "loss": 1.2332,
+ "step": 87
+ },
+ {
+ "epoch": 0.29406850459482037,
+ "grad_norm": 1.086672162070916,
+ "learning_rate": 4.944991673535803e-05,
+ "loss": 1.2241,
+ "step": 88
+ },
+ {
+ "epoch": 0.29741019214703424,
+ "grad_norm": 1.0006701473196968,
+ "learning_rate": 4.943085774526557e-05,
+ "loss": 1.2506,
+ "step": 89
+ },
+ {
+ "epoch": 0.3007518796992481,
+ "grad_norm": 1.0581177814326124,
+ "learning_rate": 4.941147798070748e-05,
+ "loss": 1.1867,
+ "step": 90
+ },
+ {
+ "epoch": 0.30409356725146197,
+ "grad_norm": 1.1670838320424346,
+ "learning_rate": 4.939177769613793e-05,
+ "loss": 1.2656,
+ "step": 91
+ },
+ {
+ "epoch": 0.30743525480367584,
+ "grad_norm": 1.0708821933487631,
+ "learning_rate": 4.937175715021947e-05,
+ "loss": 1.2476,
+ "step": 92
+ },
+ {
+ "epoch": 0.3107769423558897,
+ "grad_norm": 1.2274131496276997,
+ "learning_rate": 4.935141660581967e-05,
+ "loss": 1.2109,
+ "step": 93
+ },
+ {
+ "epoch": 0.3141186299081036,
+ "grad_norm": 1.1219369001475894,
+ "learning_rate": 4.933075633000762e-05,
+ "loss": 1.1672,
+ "step": 94
+ },
+ {
+ "epoch": 0.31746031746031744,
+ "grad_norm": 1.3306757923587944,
+ "learning_rate": 4.9309776594050485e-05,
+ "loss": 1.2768,
+ "step": 95
+ },
+ {
+ "epoch": 0.3208020050125313,
+ "grad_norm": 1.2323559759209257,
+ "learning_rate": 4.928847767340985e-05,
+ "loss": 1.2487,
+ "step": 96
+ },
+ {
+ "epoch": 0.3241436925647452,
+ "grad_norm": 1.3365381410958403,
+ "learning_rate": 4.9266859847738216e-05,
+ "loss": 1.3815,
+ "step": 97
+ },
+ {
+ "epoch": 0.32748538011695905,
+ "grad_norm": 1.0365471654338871,
+ "learning_rate": 4.9244923400875245e-05,
+ "loss": 1.2593,
+ "step": 98
+ },
+ {
+ "epoch": 0.3308270676691729,
+ "grad_norm": 1.2743177210627519,
+ "learning_rate": 4.9222668620844045e-05,
+ "loss": 1.2763,
+ "step": 99
+ },
+ {
+ "epoch": 0.3341687552213868,
+ "grad_norm": 1.134125425689655,
+ "learning_rate": 4.920009579984745e-05,
+ "loss": 1.213,
+ "step": 100
+ },
+ {
+ "epoch": 0.33751044277360065,
+ "grad_norm": 1.09970050960468,
+ "learning_rate": 4.917720523426408e-05,
+ "loss": 1.124,
+ "step": 101
+ },
+ {
+ "epoch": 0.3408521303258145,
+ "grad_norm": 1.206078654647485,
+ "learning_rate": 4.915399722464455e-05,
+ "loss": 1.2306,
+ "step": 102
+ },
+ {
+ "epoch": 0.3441938178780284,
+ "grad_norm": 1.1159637768382946,
+ "learning_rate": 4.9130472075707476e-05,
+ "loss": 1.1426,
+ "step": 103
+ },
+ {
+ "epoch": 0.34753550543024225,
+ "grad_norm": 0.9446952904809814,
+ "learning_rate": 4.910663009633544e-05,
+ "loss": 1.1786,
+ "step": 104
+ },
+ {
+ "epoch": 0.3508771929824561,
+ "grad_norm": 1.1367001934835825,
+ "learning_rate": 4.9082471599571015e-05,
+ "loss": 1.2124,
+ "step": 105
+ },
+ {
+ "epoch": 0.35421888053467,
+ "grad_norm": 1.0634444839475792,
+ "learning_rate": 4.90579969026126e-05,
+ "loss": 1.2518,
+ "step": 106
+ },
+ {
+ "epoch": 0.35756056808688386,
+ "grad_norm": 1.0063586119274412,
+ "learning_rate": 4.9033206326810254e-05,
+ "loss": 1.1361,
+ "step": 107
+ },
+ {
+ "epoch": 0.3609022556390977,
+ "grad_norm": 0.9565419392852244,
+ "learning_rate": 4.900810019766151e-05,
+ "loss": 1.2128,
+ "step": 108
+ },
+ {
+ "epoch": 0.3642439431913116,
+ "grad_norm": 0.9787104337545581,
+ "learning_rate": 4.8982678844807074e-05,
+ "loss": 1.212,
+ "step": 109
+ },
+ {
+ "epoch": 0.36758563074352546,
+ "grad_norm": 0.9780003889680408,
+ "learning_rate": 4.895694260202649e-05,
+ "loss": 1.2376,
+ "step": 110
+ },
+ {
+ "epoch": 0.37092731829573933,
+ "grad_norm": 1.0138882081963447,
+ "learning_rate": 4.8930891807233794e-05,
+ "loss": 1.2159,
+ "step": 111
+ },
+ {
+ "epoch": 0.3742690058479532,
+ "grad_norm": 0.9274261672525066,
+ "learning_rate": 4.890452680247305e-05,
+ "loss": 1.1519,
+ "step": 112
+ },
+ {
+ "epoch": 0.37761069340016706,
+ "grad_norm": 1.0636922620211422,
+ "learning_rate": 4.887784793391386e-05,
+ "loss": 1.1895,
+ "step": 113
+ },
+ {
+ "epoch": 0.38095238095238093,
+ "grad_norm": 0.9764663835059763,
+ "learning_rate": 4.885085555184681e-05,
+ "loss": 1.1014,
+ "step": 114
+ },
+ {
+ "epoch": 0.3842940685045948,
+ "grad_norm": 1.0579986191441333,
+ "learning_rate": 4.882355001067892e-05,
+ "loss": 1.2347,
+ "step": 115
+ },
+ {
+ "epoch": 0.38763575605680867,
+ "grad_norm": 0.969263591488447,
+ "learning_rate": 4.8795931668928905e-05,
+ "loss": 1.1564,
+ "step": 116
+ },
+ {
+ "epoch": 0.39097744360902253,
+ "grad_norm": 0.9798160234272053,
+ "learning_rate": 4.8768000889222565e-05,
+ "loss": 1.1676,
+ "step": 117
+ },
+ {
+ "epoch": 0.3943191311612364,
+ "grad_norm": 1.0434982392991363,
+ "learning_rate": 4.8739758038287945e-05,
+ "loss": 1.2,
+ "step": 118
+ },
+ {
+ "epoch": 0.39766081871345027,
+ "grad_norm": 0.9650348846186049,
+ "learning_rate": 4.8711203486950565e-05,
+ "loss": 1.223,
+ "step": 119
+ },
+ {
+ "epoch": 0.40100250626566414,
+ "grad_norm": 0.9624246901303686,
+ "learning_rate": 4.8682337610128526e-05,
+ "loss": 1.1558,
+ "step": 120
+ },
+ {
+ "epoch": 0.404344193817878,
+ "grad_norm": 1.1488802443868404,
+ "learning_rate": 4.8653160786827615e-05,
+ "loss": 1.2288,
+ "step": 121
+ },
+ {
+ "epoch": 0.4076858813700919,
+ "grad_norm": 1.0238637619561288,
+ "learning_rate": 4.86236734001363e-05,
+ "loss": 1.2327,
+ "step": 122
+ },
+ {
+ "epoch": 0.41102756892230574,
+ "grad_norm": 1.11719239220642,
+ "learning_rate": 4.8593875837220695e-05,
+ "loss": 1.1635,
+ "step": 123
+ },
+ {
+ "epoch": 0.4143692564745196,
+ "grad_norm": 1.009353186551115,
+ "learning_rate": 4.8563768489319526e-05,
+ "loss": 1.1562,
+ "step": 124
+ },
+ {
+ "epoch": 0.4177109440267335,
+ "grad_norm": 0.977347992720917,
+ "learning_rate": 4.8533351751738953e-05,
+ "loss": 1.2481,
+ "step": 125
+ },
+ {
+ "epoch": 0.42105263157894735,
+ "grad_norm": 0.945296219380299,
+ "learning_rate": 4.8502626023847376e-05,
+ "loss": 1.2216,
+ "step": 126
+ },
+ {
+ "epoch": 0.4243943191311612,
+ "grad_norm": 1.0795727186569568,
+ "learning_rate": 4.847159170907022e-05,
+ "loss": 1.2282,
+ "step": 127
+ },
+ {
+ "epoch": 0.4277360066833751,
+ "grad_norm": 0.9860989054801752,
+ "learning_rate": 4.844024921488459e-05,
+ "loss": 1.1639,
+ "step": 128
+ },
+ {
+ "epoch": 0.43107769423558895,
+ "grad_norm": 1.1401469693168065,
+ "learning_rate": 4.8408598952814003e-05,
+ "loss": 1.2557,
+ "step": 129
+ },
+ {
+ "epoch": 0.4344193817878028,
+ "grad_norm": 0.9621929127805111,
+ "learning_rate": 4.837664133842289e-05,
+ "loss": 1.1059,
+ "step": 130
+ },
+ {
+ "epoch": 0.4377610693400167,
+ "grad_norm": 1.1183621390483012,
+ "learning_rate": 4.834437679131121e-05,
+ "loss": 1.1732,
+ "step": 131
+ },
+ {
+ "epoch": 0.44110275689223055,
+ "grad_norm": 1.0739625747121817,
+ "learning_rate": 4.8311805735108894e-05,
+ "loss": 1.247,
+ "step": 132
+ },
+ {
+ "epoch": 0.4444444444444444,
+ "grad_norm": 1.3490735034401256,
+ "learning_rate": 4.827892859747033e-05,
+ "loss": 1.2512,
+ "step": 133
+ },
+ {
+ "epoch": 0.4477861319966583,
+ "grad_norm": 1.087730861948047,
+ "learning_rate": 4.824574581006869e-05,
+ "loss": 1.3701,
+ "step": 134
+ },
+ {
+ "epoch": 0.45112781954887216,
+ "grad_norm": 0.9963500854374578,
+ "learning_rate": 4.821225780859032e-05,
+ "loss": 1.2301,
+ "step": 135
+ },
+ {
+ "epoch": 0.454469507101086,
+ "grad_norm": 1.0799658734382207,
+ "learning_rate": 4.817846503272898e-05,
+ "loss": 1.1723,
+ "step": 136
+ },
+ {
+ "epoch": 0.4578111946532999,
+ "grad_norm": 0.990853705689885,
+ "learning_rate": 4.814436792618008e-05,
+ "loss": 1.137,
+ "step": 137
+ },
+ {
+ "epoch": 0.46115288220551376,
+ "grad_norm": 1.1057038847113485,
+ "learning_rate": 4.810996693663489e-05,
+ "loss": 1.2275,
+ "step": 138
+ },
+ {
+ "epoch": 0.4644945697577276,
+ "grad_norm": 1.060021921008829,
+ "learning_rate": 4.807526251577458e-05,
+ "loss": 1.1548,
+ "step": 139
+ },
+ {
+ "epoch": 0.4678362573099415,
+ "grad_norm": 1.0494665197923803,
+ "learning_rate": 4.8040255119264365e-05,
+ "loss": 1.1501,
+ "step": 140
+ },
+ {
+ "epoch": 0.47117794486215536,
+ "grad_norm": 0.9677143325940332,
+ "learning_rate": 4.8004945206747523e-05,
+ "loss": 1.2043,
+ "step": 141
+ },
+ {
+ "epoch": 0.47451963241436923,
+ "grad_norm": 1.1258991580511044,
+ "learning_rate": 4.796933324183928e-05,
+ "loss": 1.2874,
+ "step": 142
+ },
+ {
+ "epoch": 0.4778613199665831,
+ "grad_norm": 1.0606416431046548,
+ "learning_rate": 4.7933419692120816e-05,
+ "loss": 1.2155,
+ "step": 143
+ },
+ {
+ "epoch": 0.48120300751879697,
+ "grad_norm": 1.1260596252923112,
+ "learning_rate": 4.789720502913309e-05,
+ "loss": 1.2343,
+ "step": 144
+ },
+ {
+ "epoch": 0.48454469507101083,
+ "grad_norm": 0.9662307671780914,
+ "learning_rate": 4.78606897283706e-05,
+ "loss": 1.3106,
+ "step": 145
+ },
+ {
+ "epoch": 0.4878863826232247,
+ "grad_norm": 1.120463922826075,
+ "learning_rate": 4.782387426927524e-05,
+ "loss": 1.3122,
+ "step": 146
+ },
+ {
+ "epoch": 0.49122807017543857,
+ "grad_norm": 0.9864203109388633,
+ "learning_rate": 4.7786759135229903e-05,
+ "loss": 1.2048,
+ "step": 147
+ },
+ {
+ "epoch": 0.49456975772765244,
+ "grad_norm": 0.9891356625756446,
+ "learning_rate": 4.774934481355222e-05,
+ "loss": 1.2597,
+ "step": 148
+ },
+ {
+ "epoch": 0.4979114452798663,
+ "grad_norm": 1.024223872892248,
+ "learning_rate": 4.7711631795488096e-05,
+ "loss": 1.1611,
+ "step": 149
+ },
+ {
+ "epoch": 0.5012531328320802,
+ "grad_norm": 0.9945156162486835,
+ "learning_rate": 4.7673620576205294e-05,
+ "loss": 1.2275,
+ "step": 150
+ },
+ {
+ "epoch": 0.5012531328320802,
+ "eval_loss": 1.3914530277252197,
+ "eval_runtime": 51.6906,
+ "eval_samples_per_second": 0.426,
+ "eval_steps_per_second": 0.116,
+ "step": 150
+ },
+ {
+ "epoch": 0.504594820384294,
+ "grad_norm": 0.9560597210744347,
+ "learning_rate": 4.763531165478694e-05,
+ "loss": 1.2391,
+ "step": 151
+ },
+ {
+ "epoch": 0.5079365079365079,
+ "grad_norm": 0.948744365020844,
+ "learning_rate": 4.759670553422493e-05,
+ "loss": 1.1603,
+ "step": 152
+ },
+ {
+ "epoch": 0.5112781954887218,
+ "grad_norm": 1.0551198879801713,
+ "learning_rate": 4.755780272141338e-05,
+ "loss": 1.2125,
+ "step": 153
+ },
+ {
+ "epoch": 0.5146198830409356,
+ "grad_norm": 0.8750586318497058,
+ "learning_rate": 4.751860372714191e-05,
+ "loss": 1.1731,
+ "step": 154
+ },
+ {
+ "epoch": 0.5179615705931495,
+ "grad_norm": 1.0142475214861557,
+ "learning_rate": 4.7479109066089e-05,
+ "loss": 1.2879,
+ "step": 155
+ },
+ {
+ "epoch": 0.5213032581453634,
+ "grad_norm": 0.9748413470244546,
+ "learning_rate": 4.743931925681519e-05,
+ "loss": 1.268,
+ "step": 156
+ },
+ {
+ "epoch": 0.5246449456975772,
+ "grad_norm": 1.0082053632672576,
+ "learning_rate": 4.739923482175628e-05,
+ "loss": 1.2415,
+ "step": 157
+ },
+ {
+ "epoch": 0.5279866332497911,
+ "grad_norm": 0.9644430877325791,
+ "learning_rate": 4.735885628721648e-05,
+ "loss": 1.1902,
+ "step": 158
+ },
+ {
+ "epoch": 0.531328320802005,
+ "grad_norm": 0.9511297910602282,
+ "learning_rate": 4.73181841833615e-05,
+ "loss": 1.3144,
+ "step": 159
+ },
+ {
+ "epoch": 0.5346700083542189,
+ "grad_norm": 0.9441164162609572,
+ "learning_rate": 4.7277219044211556e-05,
+ "loss": 1.1854,
+ "step": 160
+ },
+ {
+ "epoch": 0.5380116959064327,
+ "grad_norm": 0.8932457698431419,
+ "learning_rate": 4.723596140763442e-05,
+ "loss": 1.119,
+ "step": 161
+ },
+ {
+ "epoch": 0.5413533834586466,
+ "grad_norm": 0.9849781123317826,
+ "learning_rate": 4.71944118153383e-05,
+ "loss": 1.1462,
+ "step": 162
+ },
+ {
+ "epoch": 0.5446950710108605,
+ "grad_norm": 1.1331488120142974,
+ "learning_rate": 4.7152570812864754e-05,
+ "loss": 1.2394,
+ "step": 163
+ },
+ {
+ "epoch": 0.5480367585630743,
+ "grad_norm": 0.9196323880752558,
+ "learning_rate": 4.711043894958153e-05,
+ "loss": 1.2024,
+ "step": 164
+ },
+ {
+ "epoch": 0.5513784461152882,
+ "grad_norm": 1.0274558354445025,
+ "learning_rate": 4.706801677867534e-05,
+ "loss": 1.2215,
+ "step": 165
+ },
+ {
+ "epoch": 0.5547201336675021,
+ "grad_norm": 0.9415019585312526,
+ "learning_rate": 4.702530485714461e-05,
+ "loss": 1.2155,
+ "step": 166
+ },
+ {
+ "epoch": 0.5580618212197159,
+ "grad_norm": 0.8502808406712555,
+ "learning_rate": 4.6982303745792155e-05,
+ "loss": 1.1538,
+ "step": 167
+ },
+ {
+ "epoch": 0.5614035087719298,
+ "grad_norm": 0.9941266995251048,
+ "learning_rate": 4.6939014009217816e-05,
+ "loss": 1.1087,
+ "step": 168
+ },
+ {
+ "epoch": 0.5647451963241437,
+ "grad_norm": 0.9822704957931621,
+ "learning_rate": 4.689543621581107e-05,
+ "loss": 1.249,
+ "step": 169
+ },
+ {
+ "epoch": 0.5680868838763575,
+ "grad_norm": 0.970176516153764,
+ "learning_rate": 4.68515709377435e-05,
+ "loss": 1.2191,
+ "step": 170
+ },
+ {
+ "epoch": 0.5714285714285714,
+ "grad_norm": 0.9257237748439777,
+ "learning_rate": 4.680741875096141e-05,
+ "loss": 1.1665,
+ "step": 171
+ },
+ {
+ "epoch": 0.5747702589807853,
+ "grad_norm": 0.9741770014363567,
+ "learning_rate": 4.676298023517811e-05,
+ "loss": 1.2693,
+ "step": 172
+ },
+ {
+ "epoch": 0.5781119465329991,
+ "grad_norm": 0.9190281388578779,
+ "learning_rate": 4.671825597386642e-05,
+ "loss": 1.1128,
+ "step": 173
+ },
+ {
+ "epoch": 0.581453634085213,
+ "grad_norm": 1.0223086002479052,
+ "learning_rate": 4.6673246554250945e-05,
+ "loss": 1.1661,
+ "step": 174
+ },
+ {
+ "epoch": 0.5847953216374269,
+ "grad_norm": 1.0427490752652702,
+ "learning_rate": 4.66279525673004e-05,
+ "loss": 1.2023,
+ "step": 175
+ },
+ {
+ "epoch": 0.5881370091896407,
+ "grad_norm": 0.9465248631048592,
+ "learning_rate": 4.658237460771984e-05,
+ "loss": 1.1683,
+ "step": 176
+ },
+ {
+ "epoch": 0.5914786967418546,
+ "grad_norm": 1.1640631893552786,
+ "learning_rate": 4.653651327394284e-05,
+ "loss": 1.2079,
+ "step": 177
+ },
+ {
+ "epoch": 0.5948203842940685,
+ "grad_norm": 0.9271984872996437,
+ "learning_rate": 4.6490369168123646e-05,
+ "loss": 1.1125,
+ "step": 178
+ },
+ {
+ "epoch": 0.5981620718462823,
+ "grad_norm": 0.9641228986169222,
+ "learning_rate": 4.644394289612928e-05,
+ "loss": 1.1637,
+ "step": 179
+ },
+ {
+ "epoch": 0.6015037593984962,
+ "grad_norm": 0.9782799741609496,
+ "learning_rate": 4.6397235067531564e-05,
+ "loss": 1.2689,
+ "step": 180
+ },
+ {
+ "epoch": 0.6048454469507101,
+ "grad_norm": 1.0763810726777687,
+ "learning_rate": 4.635024629559914e-05,
+ "loss": 1.2241,
+ "step": 181
+ },
+ {
+ "epoch": 0.6081871345029239,
+ "grad_norm": 0.9048533754587863,
+ "learning_rate": 4.630297719728939e-05,
+ "loss": 1.1508,
+ "step": 182
+ },
+ {
+ "epoch": 0.6115288220551378,
+ "grad_norm": 1.0817329026988984,
+ "learning_rate": 4.625542839324036e-05,
+ "loss": 1.2484,
+ "step": 183
+ },
+ {
+ "epoch": 0.6148705096073517,
+ "grad_norm": 0.9479340425567955,
+ "learning_rate": 4.6207600507762595e-05,
+ "loss": 1.1593,
+ "step": 184
+ },
+ {
+ "epoch": 0.6182121971595655,
+ "grad_norm": 0.9468319510477629,
+ "learning_rate": 4.615949416883096e-05,
+ "loss": 1.1732,
+ "step": 185
+ },
+ {
+ "epoch": 0.6215538847117794,
+ "grad_norm": 1.10744893638066,
+ "learning_rate": 4.611111000807637e-05,
+ "loss": 1.2359,
+ "step": 186
+ },
+ {
+ "epoch": 0.6248955722639933,
+ "grad_norm": 1.0229761924680858,
+ "learning_rate": 4.6062448660777517e-05,
+ "loss": 1.1828,
+ "step": 187
+ },
+ {
+ "epoch": 0.6282372598162071,
+ "grad_norm": 0.921578652654033,
+ "learning_rate": 4.601351076585253e-05,
+ "loss": 1.1131,
+ "step": 188
+ },
+ {
+ "epoch": 0.631578947368421,
+ "grad_norm": 0.9644681681801379,
+ "learning_rate": 4.596429696585054e-05,
+ "loss": 1.2319,
+ "step": 189
+ },
+ {
+ "epoch": 0.6349206349206349,
+ "grad_norm": 1.0152012900492065,
+ "learning_rate": 4.591480790694334e-05,
+ "loss": 1.229,
+ "step": 190
+ },
+ {
+ "epoch": 0.6382623224728488,
+ "grad_norm": 1.0332684046723948,
+ "learning_rate": 4.5865044238916785e-05,
+ "loss": 1.2663,
+ "step": 191
+ },
+ {
+ "epoch": 0.6416040100250626,
+ "grad_norm": 0.9861329165286253,
+ "learning_rate": 4.581500661516237e-05,
+ "loss": 1.2411,
+ "step": 192
+ },
+ {
+ "epoch": 0.6449456975772765,
+ "grad_norm": 1.008947869019847,
+ "learning_rate": 4.576469569266856e-05,
+ "loss": 1.1937,
+ "step": 193
+ },
+ {
+ "epoch": 0.6482873851294904,
+ "grad_norm": 1.0351378349442564,
+ "learning_rate": 4.57141121320122e-05,
+ "loss": 1.2424,
+ "step": 194
+ },
+ {
+ "epoch": 0.6516290726817042,
+ "grad_norm": 1.1292027980790191,
+ "learning_rate": 4.5663256597349874e-05,
+ "loss": 1.1868,
+ "step": 195
+ },
+ {
+ "epoch": 0.6549707602339181,
+ "grad_norm": 1.1325783369030435,
+ "learning_rate": 4.561212975640914e-05,
+ "loss": 1.1521,
+ "step": 196
+ },
+ {
+ "epoch": 0.658312447786132,
+ "grad_norm": 1.0325656509927605,
+ "learning_rate": 4.556073228047977e-05,
+ "loss": 1.178,
+ "step": 197
+ },
+ {
+ "epoch": 0.6616541353383458,
+ "grad_norm": 1.1093263598080985,
+ "learning_rate": 4.550906484440495e-05,
+ "loss": 1.1982,
+ "step": 198
+ },
+ {
+ "epoch": 0.6649958228905597,
+ "grad_norm": 1.114711642213788,
+ "learning_rate": 4.5457128126572423e-05,
+ "loss": 1.1487,
+ "step": 199
+ },
+ {
+ "epoch": 0.6683375104427736,
+ "grad_norm": 0.9847883166674106,
+ "learning_rate": 4.540492280890555e-05,
+ "loss": 1.2062,
+ "step": 200
+ },
+ {
+ "epoch": 0.6716791979949874,
+ "grad_norm": 1.1686727604239142,
+ "learning_rate": 4.5352449576854384e-05,
+ "loss": 1.278,
+ "step": 201
+ },
+ {
+ "epoch": 0.6750208855472013,
+ "grad_norm": 1.017923076403849,
+ "learning_rate": 4.529970911938669e-05,
+ "loss": 1.1474,
+ "step": 202
+ },
+ {
+ "epoch": 0.6783625730994152,
+ "grad_norm": 1.040704866427295,
+ "learning_rate": 4.524670212897884e-05,
+ "loss": 1.076,
+ "step": 203
+ },
+ {
+ "epoch": 0.681704260651629,
+ "grad_norm": 1.0235899496431877,
+ "learning_rate": 4.519342930160677e-05,
+ "loss": 1.2,
+ "step": 204
+ },
+ {
+ "epoch": 0.6850459482038429,
+ "grad_norm": 1.0833283292346052,
+ "learning_rate": 4.5139891336736817e-05,
+ "loss": 1.2234,
+ "step": 205
+ },
+ {
+ "epoch": 0.6883876357560568,
+ "grad_norm": 1.0293109987669598,
+ "learning_rate": 4.5086088937316565e-05,
+ "loss": 1.1714,
+ "step": 206
+ },
+ {
+ "epoch": 0.6917293233082706,
+ "grad_norm": 0.931945516283699,
+ "learning_rate": 4.503202280976557e-05,
+ "loss": 1.1223,
+ "step": 207
+ },
+ {
+ "epoch": 0.6950710108604845,
+ "grad_norm": 1.0662949673683302,
+ "learning_rate": 4.4977693663966124e-05,
+ "loss": 1.1281,
+ "step": 208
+ },
+ {
+ "epoch": 0.6984126984126984,
+ "grad_norm": 0.9541479357274659,
+ "learning_rate": 4.492310221325391e-05,
+ "loss": 1.1739,
+ "step": 209
+ },
+ {
+ "epoch": 0.7017543859649122,
+ "grad_norm": 2.071821186380696,
+ "learning_rate": 4.486824917440865e-05,
+ "loss": 1.1923,
+ "step": 210
+ },
+ {
+ "epoch": 0.7050960735171261,
+ "grad_norm": 0.946134090841913,
+ "learning_rate": 4.4813135267644693e-05,
+ "loss": 1.1303,
+ "step": 211
+ },
+ {
+ "epoch": 0.70843776106934,
+ "grad_norm": 0.9842294673980075,
+ "learning_rate": 4.475776121660156e-05,
+ "loss": 1.1622,
+ "step": 212
+ },
+ {
+ "epoch": 0.7117794486215538,
+ "grad_norm": 1.0173127525141705,
+ "learning_rate": 4.4702127748334406e-05,
+ "loss": 1.2342,
+ "step": 213
+ },
+ {
+ "epoch": 0.7151211361737677,
+ "grad_norm": 1.0630978432447922,
+ "learning_rate": 4.464623559330454e-05,
+ "loss": 1.2226,
+ "step": 214
+ },
+ {
+ "epoch": 0.7184628237259816,
+ "grad_norm": 0.8998677964481709,
+ "learning_rate": 4.4590085485369805e-05,
+ "loss": 1.1362,
+ "step": 215
+ },
+ {
+ "epoch": 0.7218045112781954,
+ "grad_norm": 0.9647726772859397,
+ "learning_rate": 4.4533678161774894e-05,
+ "loss": 1.2361,
+ "step": 216
+ },
+ {
+ "epoch": 0.7251461988304093,
+ "grad_norm": 0.9048508500249858,
+ "learning_rate": 4.447701436314176e-05,
+ "loss": 1.2976,
+ "step": 217
+ },
+ {
+ "epoch": 0.7284878863826232,
+ "grad_norm": 1.080255188574221,
+ "learning_rate": 4.4420094833459816e-05,
+ "loss": 1.2975,
+ "step": 218
+ },
+ {
+ "epoch": 0.731829573934837,
+ "grad_norm": 0.9486611908437018,
+ "learning_rate": 4.4362920320076203e-05,
+ "loss": 1.1322,
+ "step": 219
+ },
+ {
+ "epoch": 0.7351712614870509,
+ "grad_norm": 0.89345410355984,
+ "learning_rate": 4.430549157368598e-05,
+ "loss": 1.1604,
+ "step": 220
+ },
+ {
+ "epoch": 0.7385129490392648,
+ "grad_norm": 0.925519861323049,
+ "learning_rate": 4.424780934832223e-05,
+ "loss": 1.1733,
+ "step": 221
+ },
+ {
+ "epoch": 0.7418546365914787,
+ "grad_norm": 0.9646295572099951,
+ "learning_rate": 4.418987440134623e-05,
+ "loss": 1.2468,
+ "step": 222
+ },
+ {
+ "epoch": 0.7451963241436925,
+ "grad_norm": 1.0732651206685428,
+ "learning_rate": 4.4131687493437425e-05,
+ "loss": 1.3618,
+ "step": 223
+ },
+ {
+ "epoch": 0.7485380116959064,
+ "grad_norm": 0.9938198279222078,
+ "learning_rate": 4.407324938858349e-05,
+ "loss": 1.186,
+ "step": 224
+ },
+ {
+ "epoch": 0.7518796992481203,
+ "grad_norm": 0.8479511606829426,
+ "learning_rate": 4.4014560854070294e-05,
+ "loss": 1.1178,
+ "step": 225
+ },
+ {
+ "epoch": 0.7518796992481203,
+ "eval_loss": 1.3396790027618408,
+ "eval_runtime": 51.706,
+ "eval_samples_per_second": 0.425,
+ "eval_steps_per_second": 0.116,
+ "step": 225
+ },
+ {
+ "epoch": 0.7552213868003341,
+ "grad_norm": 1.0263232831005216,
+ "learning_rate": 4.395562266047182e-05,
+ "loss": 1.1985,
+ "step": 226
+ },
+ {
+ "epoch": 0.758563074352548,
+ "grad_norm": 0.8840382430386312,
+ "learning_rate": 4.389643558164005e-05,
+ "loss": 1.2017,
+ "step": 227
+ },
+ {
+ "epoch": 0.7619047619047619,
+ "grad_norm": 0.9118616467182596,
+ "learning_rate": 4.3837000394694796e-05,
+ "loss": 1.192,
+ "step": 228
+ },
+ {
+ "epoch": 0.7652464494569757,
+ "grad_norm": 0.89293130650277,
+ "learning_rate": 4.377731788001349e-05,
+ "loss": 1.1352,
+ "step": 229
+ },
+ {
+ "epoch": 0.7685881370091896,
+ "grad_norm": 0.9807040646828536,
+ "learning_rate": 4.371738882122096e-05,
+ "loss": 1.1844,
+ "step": 230
+ },
+ {
+ "epoch": 0.7719298245614035,
+ "grad_norm": 0.9931440422332504,
+ "learning_rate": 4.365721400517916e-05,
+ "loss": 1.15,
+ "step": 231
+ },
+ {
+ "epoch": 0.7752715121136173,
+ "grad_norm": 1.0024474435465867,
+ "learning_rate": 4.3596794221976754e-05,
+ "loss": 1.2061,
+ "step": 232
+ },
+ {
+ "epoch": 0.7786131996658312,
+ "grad_norm": 1.0199022805752922,
+ "learning_rate": 4.353613026491884e-05,
+ "loss": 1.2361,
+ "step": 233
+ },
+ {
+ "epoch": 0.7819548872180451,
+ "grad_norm": 1.0050111475791503,
+ "learning_rate": 4.347522293051648e-05,
+ "loss": 1.1731,
+ "step": 234
+ },
+ {
+ "epoch": 0.7852965747702589,
+ "grad_norm": 0.9751355856732374,
+ "learning_rate": 4.341407301847625e-05,
+ "loss": 1.0678,
+ "step": 235
+ },
+ {
+ "epoch": 0.7886382623224728,
+ "grad_norm": 0.9149580750612679,
+ "learning_rate": 4.335268133168974e-05,
+ "loss": 1.039,
+ "step": 236
+ },
+ {
+ "epoch": 0.7919799498746867,
+ "grad_norm": 0.96847442090601,
+ "learning_rate": 4.329104867622302e-05,
+ "loss": 1.241,
+ "step": 237
+ },
+ {
+ "epoch": 0.7953216374269005,
+ "grad_norm": 0.9528602038954359,
+ "learning_rate": 4.322917586130607e-05,
+ "loss": 1.1732,
+ "step": 238
+ },
+ {
+ "epoch": 0.7986633249791144,
+ "grad_norm": 0.9877273980590553,
+ "learning_rate": 4.3167063699322105e-05,
+ "loss": 1.1754,
+ "step": 239
+ },
+ {
+ "epoch": 0.8020050125313283,
+ "grad_norm": 0.8682490913678883,
+ "learning_rate": 4.310471300579697e-05,
+ "loss": 1.1168,
+ "step": 240
+ },
+ {
+ "epoch": 0.8053467000835421,
+ "grad_norm": 0.9028917199681886,
+ "learning_rate": 4.304212459938837e-05,
+ "loss": 1.1157,
+ "step": 241
+ },
+ {
+ "epoch": 0.808688387635756,
+ "grad_norm": 0.9101708543226708,
+ "learning_rate": 4.2979299301875184e-05,
+ "loss": 1.1691,
+ "step": 242
+ },
+ {
+ "epoch": 0.8120300751879699,
+ "grad_norm": 0.941286833245439,
+ "learning_rate": 4.291623793814663e-05,
+ "loss": 1.0917,
+ "step": 243
+ },
+ {
+ "epoch": 0.8153717627401837,
+ "grad_norm": 0.9423441770657718,
+ "learning_rate": 4.2852941336191466e-05,
+ "loss": 1.1437,
+ "step": 244
+ },
+ {
+ "epoch": 0.8187134502923976,
+ "grad_norm": 1.0019357611904858,
+ "learning_rate": 4.278941032708707e-05,
+ "loss": 1.2297,
+ "step": 245
+ },
+ {
+ "epoch": 0.8220551378446115,
+ "grad_norm": 0.9134116880628933,
+ "learning_rate": 4.27256457449886e-05,
+ "loss": 1.1763,
+ "step": 246
+ },
+ {
+ "epoch": 0.8253968253968254,
+ "grad_norm": 0.8593707359603224,
+ "learning_rate": 4.266164842711796e-05,
+ "loss": 1.097,
+ "step": 247
+ },
+ {
+ "epoch": 0.8287385129490392,
+ "grad_norm": 0.9830714729124798,
+ "learning_rate": 4.259741921375289e-05,
+ "loss": 1.1476,
+ "step": 248
+ },
+ {
+ "epoch": 0.8320802005012531,
+ "grad_norm": 1.0481652922328473,
+ "learning_rate": 4.253295894821585e-05,
+ "loss": 1.2401,
+ "step": 249
+ },
+ {
+ "epoch": 0.835421888053467,
+ "grad_norm": 0.8568238111856314,
+ "learning_rate": 4.246826847686303e-05,
+ "loss": 1.1432,
+ "step": 250
+ },
+ {
+ "epoch": 0.8387635756056808,
+ "grad_norm": 0.951730861839824,
+ "learning_rate": 4.2403348649073174e-05,
+ "loss": 1.115,
+ "step": 251
+ },
+ {
+ "epoch": 0.8421052631578947,
+ "grad_norm": 1.1229791547441785,
+ "learning_rate": 4.2338200317236445e-05,
+ "loss": 1.2719,
+ "step": 252
+ },
+ {
+ "epoch": 0.8454469507101086,
+ "grad_norm": 0.8979754621416325,
+ "learning_rate": 4.2272824336743266e-05,
+ "loss": 1.1643,
+ "step": 253
+ },
+ {
+ "epoch": 0.8487886382623224,
+ "grad_norm": 1.1038784526628656,
+ "learning_rate": 4.220722156597303e-05,
+ "loss": 1.1755,
+ "step": 254
+ },
+ {
+ "epoch": 0.8521303258145363,
+ "grad_norm": 0.8904169165772989,
+ "learning_rate": 4.214139286628289e-05,
+ "loss": 1.2078,
+ "step": 255
+ },
+ {
+ "epoch": 0.8554720133667502,
+ "grad_norm": 1.0101893594108817,
+ "learning_rate": 4.2075339101996394e-05,
+ "loss": 1.188,
+ "step": 256
+ },
+ {
+ "epoch": 0.858813700918964,
+ "grad_norm": 1.011947430714437,
+ "learning_rate": 4.200906114039219e-05,
+ "loss": 1.1815,
+ "step": 257
+ },
+ {
+ "epoch": 0.8621553884711779,
+ "grad_norm": 1.0113920659899747,
+ "learning_rate": 4.1942559851692586e-05,
+ "loss": 1.1087,
+ "step": 258
+ },
+ {
+ "epoch": 0.8654970760233918,
+ "grad_norm": 0.9436397647208052,
+ "learning_rate": 4.1875836109052166e-05,
+ "loss": 1.2688,
+ "step": 259
+ },
+ {
+ "epoch": 0.8688387635756056,
+ "grad_norm": 0.9538949648911943,
+ "learning_rate": 4.1808890788546316e-05,
+ "loss": 1.1751,
+ "step": 260
+ },
+ {
+ "epoch": 0.8721804511278195,
+ "grad_norm": 0.9935140809640558,
+ "learning_rate": 4.17417247691597e-05,
+ "loss": 1.1957,
+ "step": 261
+ },
+ {
+ "epoch": 0.8755221386800334,
+ "grad_norm": 1.0792620651192726,
+ "learning_rate": 4.1674338932774724e-05,
+ "loss": 1.1847,
+ "step": 262
+ },
+ {
+ "epoch": 0.8788638262322472,
+ "grad_norm": 1.1124795931921556,
+ "learning_rate": 4.160673416416e-05,
+ "loss": 1.1018,
+ "step": 263
+ },
+ {
+ "epoch": 0.8822055137844611,
+ "grad_norm": 0.9002781532445973,
+ "learning_rate": 4.153891135095866e-05,
+ "loss": 1.2307,
+ "step": 264
+ },
+ {
+ "epoch": 0.885547201336675,
+ "grad_norm": 0.9732414915209471,
+ "learning_rate": 4.147087138367678e-05,
+ "loss": 1.1941,
+ "step": 265
+ },
+ {
+ "epoch": 0.8888888888888888,
+ "grad_norm": 1.0519997656823685,
+ "learning_rate": 4.140261515567161e-05,
+ "loss": 1.2149,
+ "step": 266
+ },
+ {
+ "epoch": 0.8922305764411027,
+ "grad_norm": 0.9408286276236255,
+ "learning_rate": 4.133414356313988e-05,
+ "loss": 1.1152,
+ "step": 267
+ },
+ {
+ "epoch": 0.8955722639933166,
+ "grad_norm": 0.852983888206745,
+ "learning_rate": 4.126545750510605e-05,
+ "loss": 1.1358,
+ "step": 268
+ },
+ {
+ "epoch": 0.8989139515455304,
+ "grad_norm": 0.9108913540249899,
+ "learning_rate": 4.119655788341049e-05,
+ "loss": 1.0862,
+ "step": 269
+ },
+ {
+ "epoch": 0.9022556390977443,
+ "grad_norm": 1.0060408412169828,
+ "learning_rate": 4.112744560269761e-05,
+ "loss": 1.1956,
+ "step": 270
+ },
+ {
+ "epoch": 0.9055973266499582,
+ "grad_norm": 0.9488425518665664,
+ "learning_rate": 4.105812157040404e-05,
+ "loss": 1.1394,
+ "step": 271
+ },
+ {
+ "epoch": 0.908939014202172,
+ "grad_norm": 0.9505931035562328,
+ "learning_rate": 4.0988586696746674e-05,
+ "loss": 1.293,
+ "step": 272
+ },
+ {
+ "epoch": 0.9122807017543859,
+ "grad_norm": 0.9683900073885866,
+ "learning_rate": 4.0918841894710715e-05,
+ "loss": 1.2315,
+ "step": 273
+ },
+ {
+ "epoch": 0.9156223893065998,
+ "grad_norm": 1.1528253800969626,
+ "learning_rate": 4.084888808003773e-05,
+ "loss": 1.2716,
+ "step": 274
+ },
+ {
+ "epoch": 0.9189640768588136,
+ "grad_norm": 1.1069496791164177,
+ "learning_rate": 4.0778726171213576e-05,
+ "loss": 1.1983,
+ "step": 275
+ },
+ {
+ "epoch": 0.9223057644110275,
+ "grad_norm": 0.8548083007702836,
+ "learning_rate": 4.070835708945637e-05,
+ "loss": 1.1328,
+ "step": 276
+ },
+ {
+ "epoch": 0.9256474519632414,
+ "grad_norm": 0.9407270915469085,
+ "learning_rate": 4.06377817587044e-05,
+ "loss": 1.0841,
+ "step": 277
+ },
+ {
+ "epoch": 0.9289891395154553,
+ "grad_norm": 0.8954694904440085,
+ "learning_rate": 4.056700110560395e-05,
+ "loss": 1.214,
+ "step": 278
+ },
+ {
+ "epoch": 0.9323308270676691,
+ "grad_norm": 0.8961801390402134,
+ "learning_rate": 4.0496016059497204e-05,
+ "loss": 1.1721,
+ "step": 279
+ },
+ {
+ "epoch": 0.935672514619883,
+ "grad_norm": 0.9760649547793313,
+ "learning_rate": 4.042482755240997e-05,
+ "loss": 1.2666,
+ "step": 280
+ },
+ {
+ "epoch": 0.9390142021720969,
+ "grad_norm": 0.9643787134332955,
+ "learning_rate": 4.035343651903949e-05,
+ "loss": 1.1683,
+ "step": 281
+ },
+ {
+ "epoch": 0.9423558897243107,
+ "grad_norm": 0.9626450895839377,
+ "learning_rate": 4.0281843896742155e-05,
+ "loss": 1.2133,
+ "step": 282
+ },
+ {
+ "epoch": 0.9456975772765246,
+ "grad_norm": 1.2541266295867288,
+ "learning_rate": 4.021005062552119e-05,
+ "loss": 1.3502,
+ "step": 283
+ },
+ {
+ "epoch": 0.9490392648287385,
+ "grad_norm": 1.0881131031114393,
+ "learning_rate": 4.013805764801433e-05,
+ "loss": 1.3208,
+ "step": 284
+ },
+ {
+ "epoch": 0.9523809523809523,
+ "grad_norm": 0.8574920416716059,
+ "learning_rate": 4.0065865909481417e-05,
+ "loss": 1.1393,
+ "step": 285
+ },
+ {
+ "epoch": 0.9557226399331662,
+ "grad_norm": 1.0510895540058778,
+ "learning_rate": 3.9993476357792006e-05,
+ "loss": 1.117,
+ "step": 286
+ },
+ {
+ "epoch": 0.9590643274853801,
+ "grad_norm": 1.1533745861348077,
+ "learning_rate": 3.992088994341292e-05,
+ "loss": 1.2447,
+ "step": 287
+ },
+ {
+ "epoch": 0.9624060150375939,
+ "grad_norm": 1.1105061332917363,
+ "learning_rate": 3.984810761939578e-05,
+ "loss": 1.216,
+ "step": 288
+ },
+ {
+ "epoch": 0.9657477025898078,
+ "grad_norm": 0.9160630967149809,
+ "learning_rate": 3.977513034136446e-05,
+ "loss": 1.1986,
+ "step": 289
+ },
+ {
+ "epoch": 0.9690893901420217,
+ "grad_norm": 0.8445853392627862,
+ "learning_rate": 3.9701959067502564e-05,
+ "loss": 1.1836,
+ "step": 290
+ },
+ {
+ "epoch": 0.9724310776942355,
+ "grad_norm": 1.0668457057222869,
+ "learning_rate": 3.9628594758540846e-05,
+ "loss": 1.2284,
+ "step": 291
+ },
+ {
+ "epoch": 0.9757727652464494,
+ "grad_norm": 0.9615370660038071,
+ "learning_rate": 3.955503837774458e-05,
+ "loss": 1.2016,
+ "step": 292
+ },
+ {
+ "epoch": 0.9791144527986633,
+ "grad_norm": 0.9680542745347342,
+ "learning_rate": 3.948129089090093e-05,
+ "loss": 1.2164,
+ "step": 293
+ },
+ {
+ "epoch": 0.9824561403508771,
+ "grad_norm": 0.9371176041766707,
+ "learning_rate": 3.940735326630626e-05,
+ "loss": 1.1087,
+ "step": 294
+ },
+ {
+ "epoch": 0.985797827903091,
+ "grad_norm": 0.9053132575442242,
+ "learning_rate": 3.933322647475342e-05,
+ "loss": 1.0848,
+ "step": 295
+ },
+ {
+ "epoch": 0.9891395154553049,
+ "grad_norm": 0.9933354429608345,
+ "learning_rate": 3.925891148951899e-05,
+ "loss": 1.2221,
+ "step": 296
+ },
+ {
+ "epoch": 0.9924812030075187,
+ "grad_norm": 0.8955568637463743,
+ "learning_rate": 3.9184409286350535e-05,
+ "loss": 1.2599,
+ "step": 297
+ },
+ {
+ "epoch": 0.9958228905597326,
+ "grad_norm": 0.8938226719609199,
+ "learning_rate": 3.9109720843453756e-05,
+ "loss": 1.0711,
+ "step": 298
+ },
+ {
+ "epoch": 0.9991645781119465,
+ "grad_norm": 0.8770884720680571,
+ "learning_rate": 3.903484714147967e-05,
+ "loss": 1.1786,
+ "step": 299
+ },
+ {
+ "epoch": 1.0033409897682188,
+ "grad_norm": 1.0959093164885338,
+ "learning_rate": 3.895978916351172e-05,
+ "loss": 0.8299,
+ "step": 300
+ },
+ {
+ "epoch": 1.0033409897682188,
+ "eval_loss": 1.3238426446914673,
+ "eval_runtime": 12.4782,
+ "eval_samples_per_second": 1.763,
+ "eval_steps_per_second": 0.481,
+ "step": 300
+ },
+ {
+ "epoch": 1.0066819795364377,
+ "grad_norm": 0.9319113991503899,
+ "learning_rate": 3.8884547895052856e-05,
+ "loss": 0.873,
+ "step": 301
+ },
+ {
+ "epoch": 1.0100229693046565,
+ "grad_norm": 1.009627278160726,
+ "learning_rate": 3.880912432401265e-05,
+ "loss": 0.8983,
+ "step": 302
+ },
+ {
+ "epoch": 1.0133639590728754,
+ "grad_norm": 1.690445986414234,
+ "learning_rate": 3.8733519440694244e-05,
+ "loss": 0.8385,
+ "step": 303
+ },
+ {
+ "epoch": 1.0167049488410942,
+ "grad_norm": 0.875475887556019,
+ "learning_rate": 3.865773423778143e-05,
+ "loss": 0.8192,
+ "step": 304
+ },
+ {
+ "epoch": 1.020045938609313,
+ "grad_norm": 0.9324149051310555,
+ "learning_rate": 3.858176971032555e-05,
+ "loss": 0.8446,
+ "step": 305
+ },
+ {
+ "epoch": 1.0233869283775319,
+ "grad_norm": 1.0045944824797584,
+ "learning_rate": 3.850562685573246e-05,
+ "loss": 0.803,
+ "step": 306
+ },
+ {
+ "epoch": 1.0267279181457507,
+ "grad_norm": 1.0153805270432064,
+ "learning_rate": 3.8429306673749445e-05,
+ "loss": 0.8434,
+ "step": 307
+ },
+ {
+ "epoch": 1.0300689079139695,
+ "grad_norm": 0.8791912191537664,
+ "learning_rate": 3.835281016645206e-05,
+ "loss": 0.8122,
+ "step": 308
+ },
+ {
+ "epoch": 1.0334098976821884,
+ "grad_norm": 0.955967831714635,
+ "learning_rate": 3.8276138338231005e-05,
+ "loss": 0.7475,
+ "step": 309
+ },
+ {
+ "epoch": 1.0367508874504072,
+ "grad_norm": 0.9944877643459474,
+ "learning_rate": 3.819929219577894e-05,
+ "loss": 0.8238,
+ "step": 310
+ },
+ {
+ "epoch": 1.040091877218626,
+ "grad_norm": 0.9143180702052903,
+ "learning_rate": 3.81222727480772e-05,
+ "loss": 0.8181,
+ "step": 311
+ },
+ {
+ "epoch": 1.043432866986845,
+ "grad_norm": 0.8888974597110828,
+ "learning_rate": 3.8045081006382665e-05,
+ "loss": 0.8252,
+ "step": 312
+ },
+ {
+ "epoch": 1.0467738567550637,
+ "grad_norm": 0.8959954834332536,
+ "learning_rate": 3.796771798421437e-05,
+ "loss": 0.805,
+ "step": 313
+ },
+ {
+ "epoch": 1.0501148465232826,
+ "grad_norm": 0.9780665337855439,
+ "learning_rate": 3.7890184697340245e-05,
+ "loss": 0.7892,
+ "step": 314
+ },
+ {
+ "epoch": 1.0534558362915014,
+ "grad_norm": 0.8280051777925341,
+ "learning_rate": 3.781248216376381e-05,
+ "loss": 0.7809,
+ "step": 315
+ },
+ {
+ "epoch": 1.0567968260597203,
+ "grad_norm": 0.8113720404769008,
+ "learning_rate": 3.773461140371074e-05,
+ "loss": 0.7196,
+ "step": 316
+ },
+ {
+ "epoch": 1.060137815827939,
+ "grad_norm": 0.9217003877136911,
+ "learning_rate": 3.765657343961553e-05,
+ "loss": 0.7653,
+ "step": 317
+ },
+ {
+ "epoch": 1.063478805596158,
+ "grad_norm": 0.9202642653400472,
+ "learning_rate": 3.757836929610803e-05,
+ "loss": 0.7139,
+ "step": 318
+ },
+ {
+ "epoch": 1.0668197953643768,
+ "grad_norm": 0.7884990227657028,
+ "learning_rate": 3.7500000000000003e-05,
+ "loss": 0.7358,
+ "step": 319
+ },
+ {
+ "epoch": 1.0701607851325956,
+ "grad_norm": 0.873834596194201,
+ "learning_rate": 3.742146658027167e-05,
+ "loss": 0.8052,
+ "step": 320
+ },
+ {
+ "epoch": 1.0735017749008144,
+ "grad_norm": 0.9067001826221809,
+ "learning_rate": 3.7342770068058164e-05,
+ "loss": 0.701,
+ "step": 321
+ },
+ {
+ "epoch": 1.0768427646690333,
+ "grad_norm": 0.9066644745305507,
+ "learning_rate": 3.7263911496636e-05,
+ "loss": 0.8394,
+ "step": 322
+ },
+ {
+ "epoch": 1.0801837544372521,
+ "grad_norm": 0.8415263526204737,
+ "learning_rate": 3.7184891901409514e-05,
+ "loss": 0.7271,
+ "step": 323
+ },
+ {
+ "epoch": 1.083524744205471,
+ "grad_norm": 0.8351330407440325,
+ "learning_rate": 3.7105712319897265e-05,
+ "loss": 0.7699,
+ "step": 324
+ },
+ {
+ "epoch": 1.0868657339736898,
+ "grad_norm": 0.9309652326420155,
+ "learning_rate": 3.702637379171844e-05,
+ "loss": 0.7845,
+ "step": 325
+ },
+ {
+ "epoch": 1.0902067237419086,
+ "grad_norm": 0.8587868162529135,
+ "learning_rate": 3.6946877358579134e-05,
+ "loss": 0.6677,
+ "step": 326
+ },
+ {
+ "epoch": 1.0935477135101275,
+ "grad_norm": 0.8257599053579875,
+ "learning_rate": 3.6867224064258756e-05,
+ "loss": 0.7815,
+ "step": 327
+ },
+ {
+ "epoch": 1.0968887032783463,
+ "grad_norm": 0.9624273655367953,
+ "learning_rate": 3.6787414954596256e-05,
+ "loss": 0.7897,
+ "step": 328
+ },
+ {
+ "epoch": 1.1002296930465651,
+ "grad_norm": 0.944493059699902,
+ "learning_rate": 3.670745107747644e-05,
+ "loss": 0.7673,
+ "step": 329
+ },
+ {
+ "epoch": 1.1035706828147838,
+ "grad_norm": 0.929080034116846,
+ "learning_rate": 3.662733348281619e-05,
+ "loss": 0.7693,
+ "step": 330
+ },
+ {
+ "epoch": 1.1069116725830028,
+ "grad_norm": 0.8423316920482857,
+ "learning_rate": 3.6547063222550677e-05,
+ "loss": 0.7417,
+ "step": 331
+ },
+ {
+ "epoch": 1.1102526623512214,
+ "grad_norm": 0.8264975113942978,
+ "learning_rate": 3.6466641350619564e-05,
+ "loss": 0.7395,
+ "step": 332
+ },
+ {
+ "epoch": 1.1135936521194403,
+ "grad_norm": 0.8707111741842398,
+ "learning_rate": 3.638606892295315e-05,
+ "loss": 0.7883,
+ "step": 333
+ },
+ {
+ "epoch": 1.1169346418876591,
+ "grad_norm": 0.9052159861563299,
+ "learning_rate": 3.6305346997458516e-05,
+ "loss": 0.8919,
+ "step": 334
+ },
+ {
+ "epoch": 1.120275631655878,
+ "grad_norm": 0.9463481966173595,
+ "learning_rate": 3.622447663400563e-05,
+ "loss": 0.8503,
+ "step": 335
+ },
+ {
+ "epoch": 1.1236166214240968,
+ "grad_norm": 0.9451500816645809,
+ "learning_rate": 3.6143458894413465e-05,
+ "loss": 0.8774,
+ "step": 336
+ },
+ {
+ "epoch": 1.1269576111923156,
+ "grad_norm": 0.8932849188786636,
+ "learning_rate": 3.6062294842435956e-05,
+ "loss": 0.7277,
+ "step": 337
+ },
+ {
+ "epoch": 1.1302986009605345,
+ "grad_norm": 1.030017477004491,
+ "learning_rate": 3.598098554374818e-05,
+ "loss": 0.774,
+ "step": 338
+ },
+ {
+ "epoch": 1.1336395907287533,
+ "grad_norm": 0.7852586071022029,
+ "learning_rate": 3.5899532065932237e-05,
+ "loss": 0.6738,
+ "step": 339
+ },
+ {
+ "epoch": 1.1369805804969721,
+ "grad_norm": 0.944466415954939,
+ "learning_rate": 3.58179354784633e-05,
+ "loss": 0.6996,
+ "step": 340
+ },
+ {
+ "epoch": 1.140321570265191,
+ "grad_norm": 0.8970982702078042,
+ "learning_rate": 3.573619685269558e-05,
+ "loss": 0.7989,
+ "step": 341
+ },
+ {
+ "epoch": 1.1436625600334098,
+ "grad_norm": 0.8249224078826434,
+ "learning_rate": 3.565431726184818e-05,
+ "loss": 0.8146,
+ "step": 342
+ },
+ {
+ "epoch": 1.1470035498016287,
+ "grad_norm": 0.8736887097571207,
+ "learning_rate": 3.5572297780991106e-05,
+ "loss": 0.6896,
+ "step": 343
+ },
+ {
+ "epoch": 1.1503445395698475,
+ "grad_norm": 0.857438817771035,
+ "learning_rate": 3.549013948703109e-05,
+ "loss": 0.7344,
+ "step": 344
+ },
+ {
+ "epoch": 1.1536855293380663,
+ "grad_norm": 0.8839080957163322,
+ "learning_rate": 3.540784345869746e-05,
+ "loss": 0.7072,
+ "step": 345
+ },
+ {
+ "epoch": 1.1570265191062852,
+ "grad_norm": 0.8446516748750675,
+ "learning_rate": 3.532541077652799e-05,
+ "loss": 0.7722,
+ "step": 346
+ },
+ {
+ "epoch": 1.160367508874504,
+ "grad_norm": 0.8988293368576468,
+ "learning_rate": 3.5242842522854684e-05,
+ "loss": 0.7905,
+ "step": 347
+ },
+ {
+ "epoch": 1.1637084986427229,
+ "grad_norm": 0.8436031700001987,
+ "learning_rate": 3.5160139781789613e-05,
+ "loss": 0.8744,
+ "step": 348
+ },
+ {
+ "epoch": 1.1670494884109417,
+ "grad_norm": 0.8742548462166777,
+ "learning_rate": 3.5077303639210634e-05,
+ "loss": 0.8624,
+ "step": 349
+ },
+ {
+ "epoch": 1.1703904781791605,
+ "grad_norm": 0.8368319574075729,
+ "learning_rate": 3.4994335182747156e-05,
+ "loss": 0.8075,
+ "step": 350
+ },
+ {
+ "epoch": 1.1737314679473794,
+ "grad_norm": 0.8431099635101353,
+ "learning_rate": 3.491123550176586e-05,
+ "loss": 0.7896,
+ "step": 351
+ },
+ {
+ "epoch": 1.1770724577155982,
+ "grad_norm": 0.839463461544109,
+ "learning_rate": 3.482800568735638e-05,
+ "loss": 0.7758,
+ "step": 352
+ },
+ {
+ "epoch": 1.180413447483817,
+ "grad_norm": 0.8166956510001087,
+ "learning_rate": 3.474464683231698e-05,
+ "loss": 0.7709,
+ "step": 353
+ },
+ {
+ "epoch": 1.1837544372520359,
+ "grad_norm": 0.7527108106347021,
+ "learning_rate": 3.4661160031140255e-05,
+ "loss": 0.6954,
+ "step": 354
+ },
+ {
+ "epoch": 1.1870954270202547,
+ "grad_norm": 0.8885535896061523,
+ "learning_rate": 3.457754637999866e-05,
+ "loss": 0.7933,
+ "step": 355
+ },
+ {
+ "epoch": 1.1904364167884736,
+ "grad_norm": 0.7911211400323862,
+ "learning_rate": 3.4493806976730234e-05,
+ "loss": 0.7348,
+ "step": 356
+ },
+ {
+ "epoch": 1.1937774065566924,
+ "grad_norm": 0.7941043129924138,
+ "learning_rate": 3.440994292082408e-05,
+ "loss": 0.7433,
+ "step": 357
+ },
+ {
+ "epoch": 1.1971183963249112,
+ "grad_norm": 0.8629346749835708,
+ "learning_rate": 3.432595531340599e-05,
+ "loss": 0.7691,
+ "step": 358
+ },
+ {
+ "epoch": 1.20045938609313,
+ "grad_norm": 0.8076399504870541,
+ "learning_rate": 3.4241845257223995e-05,
+ "loss": 0.8387,
+ "step": 359
+ },
+ {
+ "epoch": 1.203800375861349,
+ "grad_norm": 0.8363534638021144,
+ "learning_rate": 3.4157613856633844e-05,
+ "loss": 0.7795,
+ "step": 360
+ },
+ {
+ "epoch": 1.2071413656295678,
+ "grad_norm": 0.8323776244098587,
+ "learning_rate": 3.4073262217584537e-05,
+ "loss": 0.7621,
+ "step": 361
+ },
+ {
+ "epoch": 1.2104823553977866,
+ "grad_norm": 0.8631318606885842,
+ "learning_rate": 3.3988791447603764e-05,
+ "loss": 0.7262,
+ "step": 362
+ },
+ {
+ "epoch": 1.2138233451660054,
+ "grad_norm": 0.8650278683291232,
+ "learning_rate": 3.390420265578342e-05,
+ "loss": 0.7156,
+ "step": 363
+ },
+ {
+ "epoch": 1.2171643349342243,
+ "grad_norm": 0.8210135985421421,
+ "learning_rate": 3.381949695276502e-05,
+ "loss": 0.8695,
+ "step": 364
+ },
+ {
+ "epoch": 1.220505324702443,
+ "grad_norm": 0.9816959338758507,
+ "learning_rate": 3.3734675450725075e-05,
+ "loss": 0.8165,
+ "step": 365
+ },
+ {
+ "epoch": 1.223846314470662,
+ "grad_norm": 0.8740711622449979,
+ "learning_rate": 3.364973926336057e-05,
+ "loss": 0.7659,
+ "step": 366
+ },
+ {
+ "epoch": 1.2271873042388808,
+ "grad_norm": 0.7673070294593356,
+ "learning_rate": 3.3564689505874236e-05,
+ "loss": 0.7436,
+ "step": 367
+ },
+ {
+ "epoch": 1.2305282940070996,
+ "grad_norm": 0.7912430037723751,
+ "learning_rate": 3.347952729496003e-05,
+ "loss": 0.6968,
+ "step": 368
+ },
+ {
+ "epoch": 1.2338692837753185,
+ "grad_norm": 0.7941923934515878,
+ "learning_rate": 3.3394253748788365e-05,
+ "loss": 0.8381,
+ "step": 369
+ },
+ {
+ "epoch": 1.2372102735435373,
+ "grad_norm": 0.814356027770181,
+ "learning_rate": 3.330886998699149e-05,
+ "loss": 0.7719,
+ "step": 370
+ },
+ {
+ "epoch": 1.2405512633117561,
+ "grad_norm": 0.8646739914125222,
+ "learning_rate": 3.322337713064879e-05,
+ "loss": 0.7815,
+ "step": 371
+ },
+ {
+ "epoch": 1.243892253079975,
+ "grad_norm": 0.8622517062146566,
+ "learning_rate": 3.3137776302272016e-05,
+ "loss": 0.8027,
+ "step": 372
+ },
+ {
+ "epoch": 1.2472332428481938,
+ "grad_norm": 0.9524256434298602,
+ "learning_rate": 3.3052068625790605e-05,
+ "loss": 0.8201,
+ "step": 373
+ },
+ {
+ "epoch": 1.2505742326164127,
+ "grad_norm": 0.8916856370400298,
+ "learning_rate": 3.296625522653688e-05,
+ "loss": 0.7923,
+ "step": 374
+ },
+ {
+ "epoch": 1.2539152223846315,
+ "grad_norm": 0.8558222972029906,
+ "learning_rate": 3.288033723123131e-05,
+ "loss": 0.7559,
+ "step": 375
+ },
+ {
+ "epoch": 1.2539152223846315,
+ "eval_loss": 1.369796633720398,
+ "eval_runtime": 12.5778,
+ "eval_samples_per_second": 1.749,
+ "eval_steps_per_second": 0.477,
+ "step": 375
+ }
+ ],
+ "logging_steps": 1,
+ "max_steps": 897,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 3,
+ "save_steps": 75,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.507791558504612e+17,
+ "train_batch_size": 1,
+ "trial_name": null,
+ "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6363c7fd4076c12a13237ed928a81c0f0e0926a93b924a6150ecfa586806eb92
+size 7992