Llama2-base 7B Chinese chat ver 0.01

README.md

@@ -0,0 +1,70 @@
+---
+language:
+- zh
+- en
+tags:
+- llama2
+- llama2-base
+- llama2-base-7B
+---
+# 7B Chinese Chatbot trained based on LLama2-base 7B 
+
+## Introduction
+
+在完成了[Llama2-chat 7B Chinese](https://huggingface.co/RicardoLee/Llama2-chat-Chinese-50W) 和 [Llama2-chat 13B Chinese](https://huggingface.co/RicardoLee/Llama2-chat-13B-Chinese-50W) 的训练后，我非常好奇能否直接基于Llama2-base 系列直接进行SFT训练。这也是本模型仓库的初衷。
+
+但是在实际操作中，在用了原先chat模型的LoRA训练框架后，我发现基于Llama2 base的 LoRA 训练非常难以收敛，随时处于梯度爆炸的边缘。DeepSpeed 会频繁触发reduce scale 操作，最终scale太小越界导致训练崩溃。我遍历了LR 1e-5 - 2e-4，LoRA rank \[4, 8, 64\]，LoRA Alpha \[1,4,8,16,32\]，LoRA Dropout \[0.05, 0.1\] ，Warmup Ratio \[0.01, 0.03, 0.05\]等超参数，均无法稳定训练。因此，本模型重新回归了全参数SFT训练。其难以进行LoRA训练的原因还待分析。
+
+由于网上存在使用LoRA 在英文SFT数据集上基于Llama2-base 进行SFT训练成功的样例，因此我怀疑难以训练的原因可能是扩中文词表embedding导致训练难度大幅度提升。
+
+为了方便后来人一起分析，本模型仓库特地将训练的全部loss/LR信息附在[Material](trainer_state.json)中。
+
+训练数据使用[BELLE](https://huggingface.co/BelleGroup)项目中采样的50万SFT数据进行SFT训练。
+
+After finishing the training of [Llama2-chat 7B Chinese](https://huggingface.co/RicardoLee/Llama2-chat-Chinese-50W) and [Llama2-chat 13B Chinese](https://huggingface.co/RicardoLee/Llama2-chat-13B-Chinese-50W), I am deeply intrigued by the possibility of conducting SFT (Style-Fine-Tuning) training directly based on the Llama2-base series. This is the fundamental purpose of this model repository.
+
+**However**, in real practice, I have observed that conducting LoRA training based on the Llama2 base model, within the framework of the previous Llama2-chat SFT project, presents significant challenges in achieving convergence. The gradient explosion happens in every training step and casue reducing scale operation in Deepspeed. In the end, the scale is too small and out of bounds, causing the training to crash. I have traversed LR 1e-5 - 2e-4，LoRA rank \[4, 8, 64\]，LoRA Alpha \[1,4,8,16,32\]，LoRA Dropout \[0.05, 0.1\] ，Warmup Ratio \[0.01, 0.03, 0.05\] and other hyperparameters, all of which cannot be trained stably. Therefore, this model has reverted to full-parameter SFT training. The reasons behind the difficulties encountered during LoRA training require further analysis.
+
+As there are instances online where successful LoRA training on English SFT datasets using Llama2-base has been demonstrated, I suspect that the challenge in training might be attributed to the expansion of the Chinese word embedding, resulting in a substantial increase in training difficulty.
+
+In order to facilitate collaborative analysis for future researchers, this model repository has thoughtfully appended all training-related loss/LR information in [Material](trainer_state.json).
+
+The training data is sampled from [BELLE](https://huggingface.co/BelleGroup) project, which consists of 500,000 SFT samples.
+
+## Train Detail
+
+一些训练上的细节：
+
+1. 训练框架：该模型采用全参数SFT训练，而非LoRA
+2. Tokenizer：该模型使用了Chinese-Alpaca-Plus模型的tokenizer.model。这是因为LLama2本身的tokenizer.model同LLama1是一摸一样的。因此理论上可以完全复用Chinese-LLaMa项目的tokenizer而不会产生如何错位问题。
+3. 训练参数：受限于资源，本模型只训练了1 epoch。其LR 为2e-4。Warmup ratio 为0.01。可以看到这是一个非常激进的训练，因此本模型仓库被命名为了预发布版本。未来会接着放出3 epoch版本。
+4. 训练资源：8卡V100。21个小时
+5. 训练起始的loss：参见[Material](trainer_state.json)
+6. 训练终止的loss：参见[Material](trainer_state.json)
+
+Some details in training:
+
+1. Trianing Framework: This model adopts full-parameter SFT training instead of LoRA.
+2. Tokenizer: This model utilizes the tokenizer.model from the Chinese-Alpaca-Plus model. The reason for this choice is that the tokenizer.model in LLama2 is identical to the one used in LLama1. As a result, it is theoretically feasible to entirely reuse the tokenizer from the Chinese-LLaMa project without encountering any issues related to token misalignment.
+3. Training Parameters: Constrained by limited resources, this model was trained for only 1 epoch, with a learning rate of 2e-4 and a warmup ratio of 0.01. Obviously, this is an exceedingly aggressive training schema, hence this model repository has been labeled as the 'pre-release' version. In the future, a 3-epoch version will be released subsequently for the comparison of previous Llama2-chat Chinese models.
+4. Training Resource: 8\*V100, 21 hours.
+5. Initial Loss: Please refer to [Material](trainer_state.json)
+6. Train Loss: Please refer to [Material](trainer_state.json)
+
+## Licence
+
+本仓库的模型依照 Apache-2.0 协议开源，模型的权重的使用则需要遵循LLama2[MODEL LICENCE](LICENSE)。
+
+This repository's models are open-sourced under the Apache-2.0 license, and their weight usage must adhere to LLama2 [MODEL LICENCE](LICENSE) license.
+
+## Future Work
+
+将会在近期逐步放出
+
+1. 更大SFT数据规模训练下的模型。
+2. 13B及以下的LLama2 同LLama2-chat的模型，以供大家对比。
+
+I will release the following models:
+
+1. Models trained on larger data scale.
+2. Models trained on LLama2 and LLama2-chat (under the 13B, since I only have V100), for comparison.

all_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 1.3428778648376465,
+    "eval_runtime": 2.0111,
+    "eval_samples": 100,
+    "eval_samples_per_second": 49.725,
+    "eval_steps_per_second": 1.989,
+    "perplexity": 3.830050026427415,
+    "train_loss": 1.9416432221974707,
+    "train_runtime": 74872.2082,
+    "train_samples": 500000,
+    "train_samples_per_second": 6.678,
+    "train_steps_per_second": 0.052
+}

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "RicardoLee/Llama2-base-7B-Chinese-50W-pre_release",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 2048,
+  "model_type": "llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "pad_token_id": 0,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.31.0",
+  "use_cache": true,
+  "vocab_size": 49954
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 1.0,
+    "eval_loss": 1.3428778648376465,
+    "eval_runtime": 2.0111,
+    "eval_samples": 100,
+    "eval_samples_per_second": 49.725,
+    "eval_steps_per_second": 1.989,
+    "perplexity": 3.830050026427415
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.31.0"
+}

latest

@@ -0,0 +1 @@
+global_step3906

pytorch_model-00001-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3fb0d89aa054853f5572d10c0297a29ea5a4289232837a8efaf11b1836d3126
+size 9991794123

pytorch_model-00002-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:495cc56568d3147a85eb2aba08f80b832da26ee21ca6171608830e350706d115
+size 9894802738

pytorch_model-00003-of-00003.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2620260c3425b384632c4fe17c2e1bfae8e509320fd188c37a0804fdee2e8dc6
+size 7655505659

pytorch_model.bin.index.json

@@ -0,0 +1,330 @@
+{
+  "metadata": {
+    "total_size": 27541987328
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00003-of-00003.bin",
+    "model.embed_tokens.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.0.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.1.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.10.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.11.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.11.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.11.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.11.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.11.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.11.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.11.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.11.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.12.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.13.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.14.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.15.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.16.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.17.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.18.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.19.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.2.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.20.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.20.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.21.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.input_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.mlp.down_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.mlp.up_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.post_attention_layernorm.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.22.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.23.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.23.mlp.gate_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.23.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.23.self_attn.k_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.self_attn.o_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.self_attn.q_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.self_attn.rotary_emb.inv_freq": "pytorch_model-00002-of-00003.bin",
+    "model.layers.23.self_attn.v_proj.weight": "pytorch_model-00002-of-00003.bin",
+    "model.layers.24.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.24.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.25.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.26.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.27.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.28.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.29.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.3.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.30.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.30.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.input_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.mlp.down_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.mlp.gate_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.mlp.up_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.post_attention_layernorm.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.self_attn.k_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.self_attn.o_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.self_attn.q_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.self_attn.rotary_emb.inv_freq": "pytorch_model-00003-of-00003.bin",
+    "model.layers.31.self_attn.v_proj.weight": "pytorch_model-00003-of-00003.bin",
+    "model.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.4.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.5.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.6.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.7.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.8.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.input_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.mlp.down_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.mlp.gate_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.mlp.up_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.post_attention_layernorm.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.self_attn.k_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.self_attn.o_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.self_attn.q_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.self_attn.rotary_emb.inv_freq": "pytorch_model-00001-of-00003.bin",
+    "model.layers.9.self_attn.v_proj.weight": "pytorch_model-00001-of-00003.bin",
+    "model.norm.weight": "pytorch_model-00003-of-00003.bin"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "[PAD]",
+  "unk_token": "<unk>"
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d967e855b1213a439df6c8ce2791f869c84b4f3b6cfacf22b86440b8192a2f8
+size 757972

tokenizer_config.json

@@ -0,0 +1,35 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "use_fast": true
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "train_loss": 1.9416432221974707,
+    "train_runtime": 74872.2082,
+    "train_samples": 500000,
+    "train_samples_per_second": 6.678,
+    "train_steps_per_second": 0.052
+}

trainer_state.json

@@ -0,0 +1,2395 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9998720081914757,
+  "global_step": 3906,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0,
+      "loss": 9.2076,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2e-05,
+      "loss": 9.3403,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.5e-05,
+      "loss": 7.9144,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 6.7554,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00012,
+      "loss": 5.7716,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.00017,
+      "loss": 5.0089,
+      "step": 50
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00019999947171819797,
+      "loss": 4.383,
+      "step": 60
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.00019999352861202634,
+      "loss": 4.1286,
+      "step": 70
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001999809824411913,
+      "loss": 3.7428,
+      "step": 80
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0001999618340341782,
+      "loss": 3.4558,
+      "step": 90
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00019993608465545054,
+      "loss": 3.3284,
+      "step": 100
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00019990373600536657,
+      "loss": 3.202,
+      "step": 110
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00019986479022006677,
+      "loss": 3.1471,
+      "step": 120
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.00019981924987133289,
+      "loss": 3.0477,
+      "step": 130
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00019976711796641832,
+      "loss": 2.9572,
+      "step": 140
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00019970839794784917,
+      "loss": 2.9283,
+      "step": 150
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00019964309369319722,
+      "loss": 2.882,
+      "step": 160
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.00019957120951482363,
+      "loss": 2.799,
+      "step": 170
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00019949275015959442,
+      "loss": 2.7808,
+      "step": 180
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001994077208085668,
+      "loss": 2.7251,
+      "step": 190
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.0001993161270766472,
+      "loss": 2.6977,
+      "step": 200
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 0.00019921797501222036,
+      "loss": 2.7366,
+      "step": 210
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019911327109675003,
+      "loss": 2.7007,
+      "step": 220
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019900202224435086,
+      "loss": 2.6237,
+      "step": 230
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.00019888423580133194,
+      "loss": 2.4928,
+      "step": 240
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 0.0001987599195457116,
+      "loss": 2.6081,
+      "step": 250
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019862908168670384,
+      "loss": 2.59,
+      "step": 260
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019849173086417622,
+      "loss": 2.5477,
+      "step": 270
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.0001983478761480793,
+      "loss": 2.5678,
+      "step": 280
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 0.00019819752703784777,
+      "loss": 2.4678,
+      "step": 290
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.0001980406934617731,
+      "loss": 2.4486,
+      "step": 300
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019787738577634794,
+      "loss": 2.5024,
+      "step": 310
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019770761476558223,
+      "loss": 2.5042,
+      "step": 320
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 0.00019753139164029108,
+      "loss": 2.4463,
+      "step": 330
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019734872803735444,
+      "loss": 2.4035,
+      "step": 340
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001971596360189488,
+      "loss": 2.4444,
+      "step": 350
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.0001969641280717504,
+      "loss": 2.4404,
+      "step": 360
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 0.00019676221710611093,
+      "loss": 2.4378,
+      "step": 370
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019655391645520486,
+      "loss": 2.3842,
+      "step": 380
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.000196339239874149,
+      "loss": 2.3743,
+      "step": 390
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.00019611820153909418,
+      "loss": 2.3983,
+      "step": 400
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 0.0001958908160462892,
+      "loss": 2.3224,
+      "step": 410
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.0001956570984111169,
+      "loss": 2.3779,
+      "step": 420
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00019541706406710256,
+      "loss": 2.3706,
+      "step": 430
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.000195170728864895,
+      "loss": 2.3213,
+      "step": 440
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.0001949181090712195,
+      "loss": 2.3222,
+      "step": 450
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019465922136780396,
+      "loss": 2.3018,
+      "step": 460
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019439408285027717,
+      "loss": 2.3268,
+      "step": 470
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 0.00019412271102703992,
+      "loss": 2.2956,
+      "step": 480
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019384512381810887,
+      "loss": 2.2676,
+      "step": 490
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.00019356133955393312,
+      "loss": 2.2877,
+      "step": 500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001932713769741839,
+      "loss": 2.3348,
+      "step": 510
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 0.0001929752552265169,
+      "loss": 2.3026,
+      "step": 520
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019267299386530813,
+      "loss": 2.3344,
+      "step": 530
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019236461285036233,
+      "loss": 2.258,
+      "step": 540
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.0001920501325455952,
+      "loss": 2.2424,
+      "step": 550
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 0.00019172957371768848,
+      "loss": 2.2824,
+      "step": 560
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019140295753471872,
+      "loss": 2.2618,
+      "step": 570
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.0001910703055647595,
+      "loss": 2.243,
+      "step": 580
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019073163977445696,
+      "loss": 2.2295,
+      "step": 590
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 0.00019038698252757952,
+      "loss": 2.2533,
+      "step": 600
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00019003635658354094,
+      "loss": 2.2098,
+      "step": 610
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.0001896797850958973,
+      "loss": 2.3026,
+      "step": 620
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00018931729161081835,
+      "loss": 2.19,
+      "step": 630
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 0.00018894890006553237,
+      "loss": 2.1298,
+      "step": 640
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00018857463478674552,
+      "loss": 2.1882,
+      "step": 650
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00018819452048903561,
+      "loss": 2.1378,
+      "step": 660
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.00018780858227321988,
+      "loss": 2.1886,
+      "step": 670
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 0.0001874168456246975,
+      "loss": 2.1542,
+      "step": 680
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00018701933641176676,
+      "loss": 2.1299,
+      "step": 690
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00018661608088391671,
+      "loss": 2.164,
+      "step": 700
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.0001862071056700939,
+      "loss": 2.1497,
+      "step": 710
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 0.00018579243777694387,
+      "loss": 2.1869,
+      "step": 720
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00018537210458702773,
+      "loss": 2.1317,
+      "step": 730
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00018494613385701408,
+      "loss": 2.138,
+      "step": 740
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00018451455371584603,
+      "loss": 2.1337,
+      "step": 750
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 0.00018407739266288365,
+      "loss": 2.1062,
+      "step": 760
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00018363467956602206,
+      "loss": 2.0968,
+      "step": 770
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001831864436597853,
+      "loss": 2.095,
+      "step": 780
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.00018273271454339552,
+      "loss": 2.1348,
+      "step": 790
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 0.0001822735221788186,
+      "loss": 2.0977,
+      "step": 800
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.0001818088968887857,
+      "loss": 2.1029,
+      "step": 810
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.00018133886935479057,
+      "loss": 2.1493,
+      "step": 820
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 0.0001808634706150639,
+      "loss": 2.088,
+      "step": 830
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001803827320625234,
+      "loss": 2.1212,
+      "step": 840
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00017989668544270097,
+      "loss": 2.0288,
+      "step": 850
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.0001794053628516462,
+      "loss": 2.0432,
+      "step": 860
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00017890879673380719,
+      "loss": 2.0355,
+      "step": 870
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00017840701987988772,
+      "loss": 2.0755,
+      "step": 880
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.0001779000654246823,
+      "loss": 2.0453,
+      "step": 890
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00017738796684488772,
+      "loss": 2.0582,
+      "step": 900
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 0.00017687075795689278,
+      "loss": 2.0768,
+      "step": 910
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00017634847291454503,
+      "loss": 2.091,
+      "step": 920
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.0001758211462068955,
+      "loss": 2.0577,
+      "step": 930
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00017528881265592108,
+      "loss": 2.0704,
+      "step": 940
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 0.00017475150741422528,
+      "loss": 2.0305,
+      "step": 950
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001742092659627167,
+      "loss": 2.0256,
+      "step": 960
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.0001736621241082663,
+      "loss": 2.0357,
+      "step": 970
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00017311011798134263,
+      "loss": 1.9873,
+      "step": 980
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 0.00017255328403362606,
+      "loss": 2.0248,
+      "step": 990
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00017199165903560192,
+      "loss": 1.9927,
+      "step": 1000
+    },
+    {
+      "epoch": 0.26,
+      "eval_loss": 1.7968560457229614,
+      "eval_runtime": 2.032,
+      "eval_samples_per_second": 49.213,
+      "eval_steps_per_second": 1.969,
+      "step": 1000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00017142528007413192,
+      "loss": 1.9916,
+      "step": 1010
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00017085418455000553,
+      "loss": 2.0123,
+      "step": 1020
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 0.00017027841017546998,
+      "loss": 2.0141,
+      "step": 1030
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00016969799497174005,
+      "loss": 1.976,
+      "step": 1040
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.0001691129772664873,
+      "loss": 1.9943,
+      "step": 1050
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00016852339569130905,
+      "loss": 1.9607,
+      "step": 1060
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00016792928917917755,
+      "loss": 1.9793,
+      "step": 1070
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00016733069696186868,
+      "loss": 1.988,
+      "step": 1080
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00016672765856737178,
+      "loss": 1.9143,
+      "step": 1090
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.00016612021381727887,
+      "loss": 1.9971,
+      "step": 1100
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 0.0001655084028241555,
+      "loss": 1.9675,
+      "step": 1110
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001648922659888916,
+      "loss": 2.0046,
+      "step": 1120
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00016427184399803383,
+      "loss": 2.018,
+      "step": 1130
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.0001636471778210988,
+      "loss": 1.999,
+      "step": 1140
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 0.00016301830870786742,
+      "loss": 1.9143,
+      "step": 1150
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00016238527818566138,
+      "loss": 1.9324,
+      "step": 1160
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.0001617481280566005,
+      "loss": 1.9493,
+      "step": 1170
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00016110690039484267,
+      "loss": 1.9507,
+      "step": 1180
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 0.00016046163754380514,
+      "loss": 1.9408,
+      "step": 1190
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00015981238211336873,
+      "loss": 2.0009,
+      "step": 1200
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.00015915917697706386,
+      "loss": 1.9684,
+      "step": 1210
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001585020652692394,
+      "loss": 1.9373,
+      "step": 1220
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 0.0001578410903822145,
+      "loss": 1.9038,
+      "step": 1230
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00015717629596341288,
+      "loss": 1.9065,
+      "step": 1240
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.00015650772591248085,
+      "loss": 1.9327,
+      "step": 1250
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 0.0001558354243783882,
+      "loss": 1.935,
+      "step": 1260
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.000155159435756513,
+      "loss": 1.9339,
+      "step": 1270
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00015447980468570979,
+      "loss": 1.929,
+      "step": 1280
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00015379657604536203,
+      "loss": 1.9184,
+      "step": 1290
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00015310979495241825,
+      "loss": 1.9242,
+      "step": 1300
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00015241950675841306,
+      "loss": 1.9133,
+      "step": 1310
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.0001517257570464721,
+      "loss": 1.9014,
+      "step": 1320
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00015102859162830209,
+      "loss": 1.9283,
+      "step": 1330
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 0.00015032805654116566,
+      "loss": 1.8821,
+      "step": 1340
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00014962419804484127,
+      "loss": 1.8956,
+      "step": 1350
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00014891706261856844,
+      "loss": 1.9166,
+      "step": 1360
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00014820669695797843,
+      "loss": 1.9385,
+      "step": 1370
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 0.00014749314797201084,
+      "loss": 1.9325,
+      "step": 1380
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00014677646277981593,
+      "loss": 1.8642,
+      "step": 1390
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00014605668870764293,
+      "loss": 1.8964,
+      "step": 1400
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.0001453338732857152,
+      "loss": 1.8727,
+      "step": 1410
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 0.00014460806424509132,
+      "loss": 1.8644,
+      "step": 1420
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.0001438793095145132,
+      "loss": 1.8591,
+      "step": 1430
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00014314765721724118,
+      "loss": 1.8931,
+      "step": 1440
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00014241315566787617,
+      "loss": 1.8953,
+      "step": 1450
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 0.00014167585336916926,
+      "loss": 1.8672,
+      "step": 1460
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.0001409357990088188,
+      "loss": 1.8414,
+      "step": 1470
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00014019304145625517,
+      "loss": 1.8838,
+      "step": 1480
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00013944762975941403,
+      "loss": 1.856,
+      "step": 1490
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 0.00013877453061830693,
+      "loss": 1.8715,
+      "step": 1500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00013802421179949775,
+      "loss": 1.8323,
+      "step": 1510
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00013727138205490392,
+      "loss": 1.898,
+      "step": 1520
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00013651609109757744,
+      "loss": 1.8455,
+      "step": 1530
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 0.00013575838880309623,
+      "loss": 1.8788,
+      "step": 1540
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00013499832520627076,
+      "loss": 1.8881,
+      "step": 1550
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00013423595049783974,
+      "loss": 1.8326,
+      "step": 1560
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00013347131502115616,
+      "loss": 1.845,
+      "step": 1570
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.00013270446926886252,
+      "loss": 1.8768,
+      "step": 1580
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00013193546387955672,
+      "loss": 1.8571,
+      "step": 1590
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00013116434963444815,
+      "loss": 1.8596,
+      "step": 1600
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00013039117745400426,
+      "loss": 1.8515,
+      "step": 1610
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 0.00012961599839458825,
+      "loss": 1.8281,
+      "step": 1620
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00012883886364508718,
+      "loss": 1.7872,
+      "step": 1630
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00012805982452353213,
+      "loss": 1.8333,
+      "step": 1640
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00012727893247370918,
+      "loss": 1.7989,
+      "step": 1650
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 0.00012657458799214414,
+      "loss": 1.8662,
+      "step": 1660
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.00012579031754172398,
+      "loss": 1.8667,
+      "step": 1670
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001250043440297479,
+      "loss": 1.821,
+      "step": 1680
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 0.0001242167193579139,
+      "loss": 1.8333,
+      "step": 1690
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00012342749553695423,
+      "loss": 1.8554,
+      "step": 1700
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.0001226367246832007,
+      "loss": 1.8308,
+      "step": 1710
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00012184445901514343,
+      "loss": 1.8215,
+      "step": 1720
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 0.00012105075084998242,
+      "loss": 1.8347,
+      "step": 1730
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00012025565260017291,
+      "loss": 1.7671,
+      "step": 1740
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00011945921676996417,
+      "loss": 1.8035,
+      "step": 1750
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.00011866149595193254,
+      "loss": 1.8008,
+      "step": 1760
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0001178625428235085,
+      "loss": 1.8057,
+      "step": 1770
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00011706241014349788,
+      "loss": 1.8286,
+      "step": 1780
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00011626115074859829,
+      "loss": 1.7838,
+      "step": 1790
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00011545881754990972,
+      "loss": 1.7678,
+      "step": 1800
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 0.00011465546352944083,
+      "loss": 1.8015,
+      "step": 1810
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00011385114173661003,
+      "loss": 1.773,
+      "step": 1820
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00011304590528474257,
+      "loss": 1.7528,
+      "step": 1830
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00011223980734756319,
+      "loss": 1.7651,
+      "step": 1840
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 0.00011143290115568473,
+      "loss": 1.7817,
+      "step": 1850
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00011062523999309291,
+      "loss": 1.7699,
+      "step": 1860
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00010981687719362807,
+      "loss": 1.7672,
+      "step": 1870
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00010900786613746299,
+      "loss": 1.789,
+      "step": 1880
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 0.00010819826024757807,
+      "loss": 1.7622,
+      "step": 1890
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00010738811298623348,
+      "loss": 1.7543,
+      "step": 1900
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00010657747785143882,
+      "loss": 1.7432,
+      "step": 1910
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00010576640837342036,
+      "loss": 1.7765,
+      "step": 1920
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 0.00010495495811108622,
+      "loss": 1.768,
+      "step": 1930
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00010414318064848956,
+      "loss": 1.7852,
+      "step": 1940
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001033311295912902,
+      "loss": 1.7551,
+      "step": 1950
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.0001025188585632147,
+      "loss": 1.7474,
+      "step": 1960
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 0.00010186891940623151,
+      "loss": 1.7737,
+      "step": 1970
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00010105638760647513,
+      "loss": 1.7802,
+      "step": 1980
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 0.00010024378604824765,
+      "loss": 1.7723,
+      "step": 1990
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.943116839162797e-05,
+      "loss": 1.7412,
+      "step": 2000
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 1.5315768718719482,
+      "eval_runtime": 2.0299,
+      "eval_samples_per_second": 49.264,
+      "eval_steps_per_second": 1.971,
+      "step": 2000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.869984308751394e-05,
+      "loss": 1.7468,
+      "step": 2010
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.788734267841828e-05,
+      "loss": 1.7681,
+      "step": 2020
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.707498177847988e-05,
+      "loss": 1.8109,
+      "step": 2030
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.626281403188578e-05,
+      "loss": 1.72,
+      "step": 2040
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.545089307006811e-05,
+      "loss": 1.703,
+      "step": 2050
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.463927250816272e-05,
+      "loss": 1.7624,
+      "step": 2060
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.382800594146841e-05,
+      "loss": 1.7587,
+      "step": 2070
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.301714694190808e-05,
+      "loss": 1.7375,
+      "step": 2080
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.220674905449091e-05,
+      "loss": 1.7579,
+      "step": 2090
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.139686579377649e-05,
+      "loss": 1.7396,
+      "step": 2100
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.058755064034127e-05,
+      "loss": 1.6666,
+      "step": 2110
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 8.977885703724658e-05,
+      "loss": 1.7319,
+      "step": 2120
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.897083838650984e-05,
+      "loss": 1.7387,
+      "step": 2130
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.816354804557807e-05,
+      "loss": 1.7204,
+      "step": 2140
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.743765350485347e-05,
+      "loss": 1.7183,
+      "step": 2150
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.671243090320367e-05,
+      "loss": 1.7173,
+      "step": 2160
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.590746326848647e-05,
+      "loss": 1.7185,
+      "step": 2170
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.510342623330503e-05,
+      "loss": 1.7228,
+      "step": 2180
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.430037289218072e-05,
+      "loss": 1.7542,
+      "step": 2190
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.349835627467664e-05,
+      "loss": 1.7005,
+      "step": 2200
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.269742934189604e-05,
+      "loss": 1.7211,
+      "step": 2210
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.189764498298483e-05,
+      "loss": 1.7455,
+      "step": 2220
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.109905601163912e-05,
+      "loss": 1.6729,
+      "step": 2230
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.030171516261782e-05,
+      "loss": 1.7341,
+      "step": 2240
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.950567508826012e-05,
+      "loss": 1.7286,
+      "step": 2250
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.871098835500859e-05,
+      "loss": 1.7079,
+      "step": 2260
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.791770743993817e-05,
+      "loss": 1.7001,
+      "step": 2270
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.712588472729058e-05,
+      "loss": 1.7239,
+      "step": 2280
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.633557250501531e-05,
+      "loss": 1.7032,
+      "step": 2290
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.55468229613168e-05,
+      "loss": 1.6624,
+      "step": 2300
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.475968818120798e-05,
+      "loss": 1.7258,
+      "step": 2310
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.405269046437083e-05,
+      "loss": 1.6995,
+      "step": 2320
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.342541200785587e-05,
+      "loss": 1.6715,
+      "step": 2330
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.26428964082281e-05,
+      "loss": 1.7005,
+      "step": 2340
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.186218733274769e-05,
+      "loss": 1.6575,
+      "step": 2350
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.1083336335476e-05,
+      "loss": 1.7001,
+      "step": 2360
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.030639484777641e-05,
+      "loss": 1.6679,
+      "step": 2370
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.953141417491781e-05,
+      "loss": 1.7034,
+      "step": 2380
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.875844549268706e-05,
+      "loss": 1.6804,
+      "step": 2390
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.798753984400916e-05,
+      "loss": 1.6844,
+      "step": 2400
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.721874813557699e-05,
+      "loss": 1.7038,
+      "step": 2410
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.645212113448953e-05,
+      "loss": 1.6728,
+      "step": 2420
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.568770946489948e-05,
+      "loss": 1.682,
+      "step": 2430
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.492556360467025e-05,
+      "loss": 1.6799,
+      "step": 2440
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.416573388204282e-05,
+      "loss": 1.66,
+      "step": 2450
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.340827047231211e-05,
+      "loss": 1.6806,
+      "step": 2460
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.265322339451376e-05,
+      "loss": 1.6661,
+      "step": 2470
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.190064250812124e-05,
+      "loss": 1.6696,
+      "step": 2480
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.115057750975312e-05,
+      "loss": 1.6153,
+      "step": 2490
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.040307792989157e-05,
+      "loss": 1.6824,
+      "step": 2500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.9658193129611604e-05,
+      "loss": 1.6886,
+      "step": 2510
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.891597229732135e-05,
+      "loss": 1.6358,
+      "step": 2520
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.8176464445514166e-05,
+      "loss": 1.6462,
+      "step": 2530
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7439718407531906e-05,
+      "loss": 1.6434,
+      "step": 2540
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.670578283434016e-05,
+      "loss": 1.6459,
+      "step": 2550
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5974706191315884e-05,
+      "loss": 1.6705,
+      "step": 2560
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5246536755046706e-05,
+      "loss": 1.6638,
+      "step": 2570
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.452132261014304e-05,
+      "loss": 1.6656,
+      "step": 2580
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.379911164606304e-05,
+      "loss": 1.6572,
+      "step": 2590
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.315172891887351e-05,
+      "loss": 1.643,
+      "step": 2600
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2435355221012797e-05,
+      "loss": 1.6544,
+      "step": 2610
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.172212245066537e-05,
+      "loss": 1.628,
+      "step": 2620
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.1012077706100125e-05,
+      "loss": 1.6378,
+      "step": 2630
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.0305267875065087e-05,
+      "loss": 1.6475,
+      "step": 2640
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.9601739631690836e-05,
+      "loss": 1.5959,
+      "step": 2650
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.897140837169796e-05,
+      "loss": 1.657,
+      "step": 2660
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.827424295352793e-05,
+      "loss": 1.6716,
+      "step": 2670
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.758049324158693e-05,
+      "loss": 1.59,
+      "step": 2680
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.6890205047581745e-05,
+      "loss": 1.6442,
+      "step": 2690
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.6203423954637995e-05,
+      "loss": 1.6152,
+      "step": 2700
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.552019531429019e-05,
+      "loss": 1.6446,
+      "step": 2710
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.484056424348703e-05,
+      "loss": 1.6216,
+      "step": 2720
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.416457562161184e-05,
+      "loss": 1.6534,
+      "step": 2730
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.349227408751919e-05,
+      "loss": 1.6474,
+      "step": 2740
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.282370403658717e-05,
+      "loss": 1.6338,
+      "step": 2750
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.2158909617785525e-05,
+      "loss": 1.6473,
+      "step": 2760
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.149793473076058e-05,
+      "loss": 1.6315,
+      "step": 2770
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.084082302293617e-05,
+      "loss": 1.6516,
+      "step": 2780
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.018761788663127e-05,
+      "loss": 1.6112,
+      "step": 2790
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.953836245619488e-05,
+      "loss": 1.6077,
+      "step": 2800
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.889309960515738e-05,
+      "loss": 1.6182,
+      "step": 2810
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.82518719433995e-05,
+      "loss": 1.6072,
+      "step": 2820
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.761472181433865e-05,
+      "loss": 1.6062,
+      "step": 2830
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6981691292132604e-05,
+      "loss": 1.6332,
+      "step": 2840
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6352822178901235e-05,
+      "loss": 1.6393,
+      "step": 2850
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5728156001966154e-05,
+      "loss": 1.6401,
+      "step": 2860
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5169584051980575e-05,
+      "loss": 1.6234,
+      "step": 2870
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.461447977339909e-05,
+      "loss": 1.5814,
+      "step": 2880
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4001800370596834e-05,
+      "loss": 1.6018,
+      "step": 2890
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.339347915362796e-05,
+      "loss": 1.6172,
+      "step": 2900
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.278955629293534e-05,
+      "loss": 1.6042,
+      "step": 2910
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.219007166851673e-05,
+      "loss": 1.6119,
+      "step": 2920
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1595064867291394e-05,
+      "loss": 1.621,
+      "step": 2930
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1004575180485885e-05,
+      "loss": 1.6046,
+      "step": 2940
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.0418641601039366e-05,
+      "loss": 1.5811,
+      "step": 2950
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9837302821028956e-05,
+      "loss": 1.5635,
+      "step": 2960
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.926059722911447e-05,
+      "loss": 1.6193,
+      "step": 2970
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.86885629080035e-05,
+      "loss": 1.6067,
+      "step": 2980
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.823432416081132e-05,
+      "loss": 1.5795,
+      "step": 2990
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7670793109350358e-05,
+      "loss": 1.5891,
+      "step": 3000
+    },
+    {
+      "epoch": 0.77,
+      "eval_loss": 1.3908636569976807,
+      "eval_runtime": 2.0291,
+      "eval_samples_per_second": 49.282,
+      "eval_steps_per_second": 1.971,
+      "step": 3000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7112038311280828e-05,
+      "loss": 1.599,
+      "step": 3010
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.655809666393112e-05,
+      "loss": 1.5877,
+      "step": 3020
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.600900474679364e-05,
+      "loss": 1.6096,
+      "step": 3030
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.546479881910918e-05,
+      "loss": 1.6317,
+      "step": 3040
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4925514817472618e-05,
+      "loss": 1.5218,
+      "step": 3050
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4391188353459925e-05,
+      "loss": 1.5526,
+      "step": 3060
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3861854711276378e-05,
+      "loss": 1.5753,
+      "step": 3070
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.333754884542667e-05,
+      "loss": 1.6214,
+      "step": 3080
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.281830537840678e-05,
+      "loss": 1.591,
+      "step": 3090
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2355342955230186e-05,
+      "loss": 1.5578,
+      "step": 3100
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.189653361595686e-05,
+      "loss": 1.5684,
+      "step": 3110
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1391646203159456e-05,
+      "loss": 1.5654,
+      "step": 3120
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.089194968671713e-05,
+      "loss": 1.5803,
+      "step": 3130
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.039747706404943e-05,
+      "loss": 1.5737,
+      "step": 3140
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.99082609876164e-05,
+      "loss": 1.5444,
+      "step": 3150
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9472487573431274e-05,
+      "loss": 1.5995,
+      "step": 3160
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8993347647457706e-05,
+      "loss": 1.5803,
+      "step": 3170
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8519556989292508e-05,
+      "loss": 1.5892,
+      "step": 3180
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.8051146885663938e-05,
+      "loss": 1.6006,
+      "step": 3190
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7588148267995695e-05,
+      "loss": 1.567,
+      "step": 3200
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7130591710364486e-05,
+      "loss": 1.5557,
+      "step": 3210
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6678507427480983e-05,
+      "loss": 1.5794,
+      "step": 3220
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.6231925272694615e-05,
+      "loss": 1.5858,
+      "step": 3230
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5790874736022287e-05,
+      "loss": 1.5791,
+      "step": 3240
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.535538494220089e-05,
+      "loss": 1.5721,
+      "step": 3250
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4925484648764131e-05,
+      "loss": 1.5537,
+      "step": 3260
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.450120224414352e-05,
+      "loss": 1.5698,
+      "step": 3270
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.4082565745793686e-05,
+      "loss": 1.5529,
+      "step": 3280
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3669602798342296e-05,
+      "loss": 1.5702,
+      "step": 3290
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3262340671764584e-05,
+      "loss": 1.5273,
+      "step": 3300
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2860806259582492e-05,
+      "loss": 1.5401,
+      "step": 3310
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2504344407159785e-05,
+      "loss": 1.5753,
+      "step": 3320
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2113765387943211e-05,
+      "loss": 1.5564,
+      "step": 3330
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.172898992919923e-05,
+      "loss": 1.5189,
+      "step": 3340
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.1350043439544521e-05,
+      "loss": 1.5607,
+      "step": 3350
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0976950942680197e-05,
+      "loss": 1.539,
+      "step": 3360
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0609737075739412e-05,
+      "loss": 1.5593,
+      "step": 3370
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.0248426087660557e-05,
+      "loss": 1.5345,
+      "step": 3380
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.89304183758577e-06,
+      "loss": 1.5988,
+      "step": 3390
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.543607793285626e-06,
+      "loss": 1.5306,
+      "step": 3400
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.200147029609264e-06,
+      "loss": 1.5702,
+      "step": 3410
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.896158250762244e-06,
+      "loss": 1.5378,
+      "step": 3420
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.59704246528129e-06,
+      "loss": 1.5693,
+      "step": 3430
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.270426282311539e-06,
+      "loss": 1.5517,
+      "step": 3440
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.949867454404824e-06,
+      "loss": 1.5576,
+      "step": 3450
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 7.635387149637685e-06,
+      "loss": 1.5763,
+      "step": 3460
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 7.327006134691883e-06,
+      "loss": 1.5768,
+      "step": 3470
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 7.024744773483105e-06,
+      "loss": 1.5393,
+      "step": 3480
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.7286230258161385e-06,
+      "loss": 1.5617,
+      "step": 3490
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 6.438660446066891e-06,
+      "loss": 1.5404,
+      "step": 3500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 6.154876181891145e-06,
+      "loss": 1.5765,
+      "step": 3510
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.877288972960071e-06,
+      "loss": 1.5942,
+      "step": 3520
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.632774125747675e-06,
+      "loss": 1.5557,
+      "step": 3530
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.367011482971008e-06,
+      "loss": 1.5438,
+      "step": 3540
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 5.107497922021364e-06,
+      "loss": 1.5351,
+      "step": 3550
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.854250579856034e-06,
+      "loss": 1.5304,
+      "step": 3560
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.6072861796429665e-06,
+      "loss": 1.554,
+      "step": 3570
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.366621029656582e-06,
+      "loss": 1.5185,
+      "step": 3580
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.1322710222008065e-06,
+      "loss": 1.5746,
+      "step": 3590
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.904251632559652e-06,
+      "loss": 1.5413,
+      "step": 3600
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.6825779179752716e-06,
+      "loss": 1.5533,
+      "step": 3610
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.467264516653668e-06,
+      "loss": 1.5432,
+      "step": 3620
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.2583256467980773e-06,
+      "loss": 1.5869,
+      "step": 3630
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.055775105670056e-06,
+      "loss": 1.5374,
+      "step": 3640
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8596262686783837e-06,
+      "loss": 1.5425,
+      "step": 3650
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.6698920884958177e-06,
+      "loss": 1.5906,
+      "step": 3660
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.486585094203786e-06,
+      "loss": 1.5787,
+      "step": 3670
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.309717390464983e-06,
+      "loss": 1.5579,
+      "step": 3680
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1393006567240635e-06,
+      "loss": 1.5391,
+      "step": 3690
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9753461464364408e-06,
+      "loss": 1.5478,
+      "step": 3700
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.8178646863250548e-06,
+      "loss": 1.5777,
+      "step": 3710
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.6668666756655572e-06,
+      "loss": 1.5239,
+      "step": 3720
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5365200653588708e-06,
+      "loss": 1.4992,
+      "step": 3730
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.397867724769042e-06,
+      "loss": 1.5272,
+      "step": 3740
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.2657265680968589e-06,
+      "loss": 1.541,
+      "step": 3750
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1523738102167225e-06,
+      "loss": 1.5219,
+      "step": 3760
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.044308247886483e-06,
+      "loss": 1.5524,
+      "step": 3770
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.30443453495422e-07,
+      "loss": 1.5508,
+      "step": 3780
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 8.231207093463699e-07,
+      "loss": 1.5758,
+      "step": 3790
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.223471024881412e-07,
+      "loss": 1.5658,
+      "step": 3800
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 6.281292874978029e-07,
+      "loss": 1.5232,
+      "step": 3810
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 5.404734860412375e-07,
+      "loss": 1.5646,
+      "step": 3820
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.5938548646227154e-07,
+      "loss": 1.5771,
+      "step": 3830
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.8487064340047006e-07,
+      "loss": 1.5611,
+      "step": 3840
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.16933877437553e-07,
+      "loss": 1.6229,
+      "step": 3850
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.555796747724104e-07,
+      "loss": 1.5496,
+      "step": 3860
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.0081208692490638e-07,
+      "loss": 1.5312,
+      "step": 3870
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.5263473046833732e-07,
+      "loss": 1.5681,
+      "step": 3880
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.1105078679056747e-07,
+      "loss": 1.5128,
+      "step": 3890
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.606300188400805e-08,
+      "loss": 1.5764,
+      "step": 3900
+    },
+    {
+      "epoch": 1.0,
+      "step": 3906,
+      "total_flos": 331952415375360.0,
+      "train_loss": 1.9416432221974707,
+      "train_runtime": 74872.2082,
+      "train_samples_per_second": 6.678,
+      "train_steps_per_second": 0.052
+    }
+  ],
+  "max_steps": 3906,
+  "num_train_epochs": 1,
+  "total_flos": 331952415375360.0,
+  "trial_name": null,
+  "trial_params": null
+}