Model save
Browse files- README.md +58 -0
- all_results.json +8 -0
- generation_config.json +14 -0
- train_results.json +8 -0
- trainer_state.json +3556 -0
README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-1.5B-Instruct
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen2.5-1.5B-Open-R1-Distill-3-epoch
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- sft
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen2.5-1.5B-Open-R1-Distill-3-epoch
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="HectorHe/Qwen2.5-1.5B-Open-R1-Distill-3-epoch", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/hector_-carnegie-mellon-university/huggingface/runs/kwnp5nsh)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with SFT.
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.18.0.dev0
|
| 38 |
+
- Transformers: 4.52.0.dev0
|
| 39 |
+
- Pytorch: 2.6.0
|
| 40 |
+
- Datasets: 3.6.0
|
| 41 |
+
- Tokenizers: 0.21.2
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
Cite TRL as:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@misc{vonwerra2022trl,
|
| 51 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 52 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
|
| 53 |
+
year = 2020,
|
| 54 |
+
journal = {GitHub repository},
|
| 55 |
+
publisher = {GitHub},
|
| 56 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 57 |
+
}
|
| 58 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 3438830624964608.0,
|
| 3 |
+
"train_loss": 0.4748283546585666,
|
| 4 |
+
"train_runtime": 17138.2816,
|
| 5 |
+
"train_samples": 93733,
|
| 6 |
+
"train_samples_per_second": 16.408,
|
| 7 |
+
"train_steps_per_second": 0.128
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.1,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.52.0.dev0"
|
| 14 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 3438830624964608.0,
|
| 3 |
+
"train_loss": 0.4748283546585666,
|
| 4 |
+
"train_runtime": 17138.2816,
|
| 5 |
+
"train_samples": 93733,
|
| 6 |
+
"train_samples_per_second": 16.408,
|
| 7 |
+
"train_steps_per_second": 0.128
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3556 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2199,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0068212824010914054,
|
| 14 |
+
"grad_norm": 2.532045957962293,
|
| 15 |
+
"learning_rate": 1.818181818181818e-06,
|
| 16 |
+
"loss": 0.8581,
|
| 17 |
+
"num_tokens": 3759146.0,
|
| 18 |
+
"step": 5
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 0.013642564802182811,
|
| 22 |
+
"grad_norm": 1.5090606444318968,
|
| 23 |
+
"learning_rate": 4.0909090909090915e-06,
|
| 24 |
+
"loss": 0.8434,
|
| 25 |
+
"num_tokens": 7668808.0,
|
| 26 |
+
"step": 10
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.020463847203274217,
|
| 30 |
+
"grad_norm": 1.4089375644452828,
|
| 31 |
+
"learning_rate": 6.363636363636363e-06,
|
| 32 |
+
"loss": 0.7812,
|
| 33 |
+
"num_tokens": 11368873.0,
|
| 34 |
+
"step": 15
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 0.027285129604365622,
|
| 38 |
+
"grad_norm": 0.87029907931035,
|
| 39 |
+
"learning_rate": 8.636363636363637e-06,
|
| 40 |
+
"loss": 0.7301,
|
| 41 |
+
"num_tokens": 15118063.0,
|
| 42 |
+
"step": 20
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 0.034106412005457026,
|
| 46 |
+
"grad_norm": 0.6304280306164425,
|
| 47 |
+
"learning_rate": 1.0909090909090909e-05,
|
| 48 |
+
"loss": 0.6951,
|
| 49 |
+
"num_tokens": 18906839.0,
|
| 50 |
+
"step": 25
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.040927694406548434,
|
| 54 |
+
"grad_norm": 0.4866174158862162,
|
| 55 |
+
"learning_rate": 1.318181818181818e-05,
|
| 56 |
+
"loss": 0.6665,
|
| 57 |
+
"num_tokens": 22641755.0,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.047748976807639835,
|
| 62 |
+
"grad_norm": 0.3925544859147944,
|
| 63 |
+
"learning_rate": 1.5454545454545454e-05,
|
| 64 |
+
"loss": 0.6591,
|
| 65 |
+
"num_tokens": 26636629.0,
|
| 66 |
+
"step": 35
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.054570259208731244,
|
| 70 |
+
"grad_norm": 0.3345987178858572,
|
| 71 |
+
"learning_rate": 1.772727272727273e-05,
|
| 72 |
+
"loss": 0.6254,
|
| 73 |
+
"num_tokens": 30417967.0,
|
| 74 |
+
"step": 40
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.061391541609822645,
|
| 78 |
+
"grad_norm": 0.3135639706364999,
|
| 79 |
+
"learning_rate": 2e-05,
|
| 80 |
+
"loss": 0.626,
|
| 81 |
+
"num_tokens": 34231333.0,
|
| 82 |
+
"step": 45
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 0.06821282401091405,
|
| 86 |
+
"grad_norm": 0.2862092724525678,
|
| 87 |
+
"learning_rate": 2.2272727272727274e-05,
|
| 88 |
+
"loss": 0.6243,
|
| 89 |
+
"num_tokens": 37961424.0,
|
| 90 |
+
"step": 50
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.07503410641200546,
|
| 94 |
+
"grad_norm": 0.3002457975219281,
|
| 95 |
+
"learning_rate": 2.4545454545454545e-05,
|
| 96 |
+
"loss": 0.6224,
|
| 97 |
+
"num_tokens": 41826860.0,
|
| 98 |
+
"step": 55
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.08185538881309687,
|
| 102 |
+
"grad_norm": 0.32826550897955076,
|
| 103 |
+
"learning_rate": 2.681818181818182e-05,
|
| 104 |
+
"loss": 0.608,
|
| 105 |
+
"num_tokens": 45543403.0,
|
| 106 |
+
"step": 60
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 0.08867667121418826,
|
| 110 |
+
"grad_norm": 0.3197398874021795,
|
| 111 |
+
"learning_rate": 2.909090909090909e-05,
|
| 112 |
+
"loss": 0.6103,
|
| 113 |
+
"num_tokens": 49369486.0,
|
| 114 |
+
"step": 65
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.09549795361527967,
|
| 118 |
+
"grad_norm": 0.33920714651377604,
|
| 119 |
+
"learning_rate": 3.1363636363636365e-05,
|
| 120 |
+
"loss": 0.5984,
|
| 121 |
+
"num_tokens": 53010874.0,
|
| 122 |
+
"step": 70
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.10231923601637108,
|
| 126 |
+
"grad_norm": 0.4148929626561891,
|
| 127 |
+
"learning_rate": 3.3636363636363636e-05,
|
| 128 |
+
"loss": 0.5999,
|
| 129 |
+
"num_tokens": 56909889.0,
|
| 130 |
+
"step": 75
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.10914051841746249,
|
| 134 |
+
"grad_norm": 0.4745249491489932,
|
| 135 |
+
"learning_rate": 3.590909090909091e-05,
|
| 136 |
+
"loss": 0.5824,
|
| 137 |
+
"num_tokens": 60650570.0,
|
| 138 |
+
"step": 80
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 0.11596180081855388,
|
| 142 |
+
"grad_norm": 0.5183539226015511,
|
| 143 |
+
"learning_rate": 3.818181818181819e-05,
|
| 144 |
+
"loss": 0.5755,
|
| 145 |
+
"num_tokens": 64564660.0,
|
| 146 |
+
"step": 85
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.12278308321964529,
|
| 150 |
+
"grad_norm": 0.5504513108051802,
|
| 151 |
+
"learning_rate": 4.045454545454546e-05,
|
| 152 |
+
"loss": 0.5928,
|
| 153 |
+
"num_tokens": 68426882.0,
|
| 154 |
+
"step": 90
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.1296043656207367,
|
| 158 |
+
"grad_norm": 0.46351356469226884,
|
| 159 |
+
"learning_rate": 4.2727272727272724e-05,
|
| 160 |
+
"loss": 0.5878,
|
| 161 |
+
"num_tokens": 72252819.0,
|
| 162 |
+
"step": 95
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 0.1364256480218281,
|
| 166 |
+
"grad_norm": 0.4664664837842009,
|
| 167 |
+
"learning_rate": 4.5e-05,
|
| 168 |
+
"loss": 0.5737,
|
| 169 |
+
"num_tokens": 76160914.0,
|
| 170 |
+
"step": 100
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.1432469304229195,
|
| 174 |
+
"grad_norm": 0.4309398293439996,
|
| 175 |
+
"learning_rate": 4.7272727272727275e-05,
|
| 176 |
+
"loss": 0.5752,
|
| 177 |
+
"num_tokens": 80152955.0,
|
| 178 |
+
"step": 105
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.15006821282401092,
|
| 182 |
+
"grad_norm": 0.3564559897881336,
|
| 183 |
+
"learning_rate": 4.9545454545454553e-05,
|
| 184 |
+
"loss": 0.5782,
|
| 185 |
+
"num_tokens": 83901702.0,
|
| 186 |
+
"step": 110
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.15688949522510232,
|
| 190 |
+
"grad_norm": 0.4251960297464194,
|
| 191 |
+
"learning_rate": 4.99995929066005e-05,
|
| 192 |
+
"loss": 0.5767,
|
| 193 |
+
"num_tokens": 87691018.0,
|
| 194 |
+
"step": 115
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.16371077762619374,
|
| 198 |
+
"grad_norm": 0.4057834381063162,
|
| 199 |
+
"learning_rate": 4.9997939114912153e-05,
|
| 200 |
+
"loss": 0.5658,
|
| 201 |
+
"num_tokens": 91542428.0,
|
| 202 |
+
"step": 120
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.17053206002728513,
|
| 206 |
+
"grad_norm": 0.3965112414949712,
|
| 207 |
+
"learning_rate": 4.999501327503205e-05,
|
| 208 |
+
"loss": 0.5719,
|
| 209 |
+
"num_tokens": 95438170.0,
|
| 210 |
+
"step": 125
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.17735334242837653,
|
| 214 |
+
"grad_norm": 0.43723602405301004,
|
| 215 |
+
"learning_rate": 4.999081555238907e-05,
|
| 216 |
+
"loss": 0.5669,
|
| 217 |
+
"num_tokens": 99383692.0,
|
| 218 |
+
"step": 130
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.18417462482946795,
|
| 222 |
+
"grad_norm": 0.4642547524939693,
|
| 223 |
+
"learning_rate": 4.99853461843252e-05,
|
| 224 |
+
"loss": 0.5665,
|
| 225 |
+
"num_tokens": 103223537.0,
|
| 226 |
+
"step": 135
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.19099590723055934,
|
| 230 |
+
"grad_norm": 0.4882526595367141,
|
| 231 |
+
"learning_rate": 4.997860548008209e-05,
|
| 232 |
+
"loss": 0.5794,
|
| 233 |
+
"num_tokens": 106901687.0,
|
| 234 |
+
"step": 140
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.19781718963165076,
|
| 238 |
+
"grad_norm": 0.3270132768432918,
|
| 239 |
+
"learning_rate": 4.997059382078357e-05,
|
| 240 |
+
"loss": 0.5511,
|
| 241 |
+
"num_tokens": 110840758.0,
|
| 242 |
+
"step": 145
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.20463847203274216,
|
| 246 |
+
"grad_norm": 0.36660863914844466,
|
| 247 |
+
"learning_rate": 4.9961311659414095e-05,
|
| 248 |
+
"loss": 0.5663,
|
| 249 |
+
"num_tokens": 114521504.0,
|
| 250 |
+
"step": 150
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.21145975443383355,
|
| 254 |
+
"grad_norm": 0.41353953388884024,
|
| 255 |
+
"learning_rate": 4.995075952079313e-05,
|
| 256 |
+
"loss": 0.5503,
|
| 257 |
+
"num_tokens": 118445759.0,
|
| 258 |
+
"step": 155
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 0.21828103683492497,
|
| 262 |
+
"grad_norm": 0.39374926790537124,
|
| 263 |
+
"learning_rate": 4.9938938001545504e-05,
|
| 264 |
+
"loss": 0.5588,
|
| 265 |
+
"num_tokens": 122311693.0,
|
| 266 |
+
"step": 160
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.22510231923601637,
|
| 270 |
+
"grad_norm": 0.3963624580332336,
|
| 271 |
+
"learning_rate": 4.992584777006765e-05,
|
| 272 |
+
"loss": 0.5614,
|
| 273 |
+
"num_tokens": 126094524.0,
|
| 274 |
+
"step": 165
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.23192360163710776,
|
| 278 |
+
"grad_norm": 0.41321778807677423,
|
| 279 |
+
"learning_rate": 4.99114895664898e-05,
|
| 280 |
+
"loss": 0.5547,
|
| 281 |
+
"num_tokens": 129971056.0,
|
| 282 |
+
"step": 170
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.23874488403819918,
|
| 286 |
+
"grad_norm": 0.4592368771120155,
|
| 287 |
+
"learning_rate": 4.9895864202634194e-05,
|
| 288 |
+
"loss": 0.5695,
|
| 289 |
+
"num_tokens": 133765982.0,
|
| 290 |
+
"step": 175
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.24556616643929058,
|
| 294 |
+
"grad_norm": 0.5030729658739199,
|
| 295 |
+
"learning_rate": 4.987897256196912e-05,
|
| 296 |
+
"loss": 0.5494,
|
| 297 |
+
"num_tokens": 137522281.0,
|
| 298 |
+
"step": 180
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.252387448840382,
|
| 302 |
+
"grad_norm": 0.37507479857819553,
|
| 303 |
+
"learning_rate": 4.986081559955899e-05,
|
| 304 |
+
"loss": 0.5437,
|
| 305 |
+
"num_tokens": 141196084.0,
|
| 306 |
+
"step": 185
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 0.2592087312414734,
|
| 310 |
+
"grad_norm": 0.3840278827691952,
|
| 311 |
+
"learning_rate": 4.984139434201032e-05,
|
| 312 |
+
"loss": 0.5547,
|
| 313 |
+
"num_tokens": 145136704.0,
|
| 314 |
+
"step": 190
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.2660300136425648,
|
| 318 |
+
"grad_norm": 0.3714874079416638,
|
| 319 |
+
"learning_rate": 4.982070988741372e-05,
|
| 320 |
+
"loss": 0.5425,
|
| 321 |
+
"num_tokens": 148940122.0,
|
| 322 |
+
"step": 195
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 0.2728512960436562,
|
| 326 |
+
"grad_norm": 0.3818119404771569,
|
| 327 |
+
"learning_rate": 4.979876340528179e-05,
|
| 328 |
+
"loss": 0.5521,
|
| 329 |
+
"num_tokens": 152809678.0,
|
| 330 |
+
"step": 200
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.27967257844474763,
|
| 334 |
+
"grad_norm": 0.44357408734230386,
|
| 335 |
+
"learning_rate": 4.977555613648297e-05,
|
| 336 |
+
"loss": 0.5456,
|
| 337 |
+
"num_tokens": 156683573.0,
|
| 338 |
+
"step": 205
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.286493860845839,
|
| 342 |
+
"grad_norm": 0.4239986360553216,
|
| 343 |
+
"learning_rate": 4.975108939317143e-05,
|
| 344 |
+
"loss": 0.5469,
|
| 345 |
+
"num_tokens": 160433326.0,
|
| 346 |
+
"step": 210
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.2933151432469304,
|
| 350 |
+
"grad_norm": 0.31362423852071797,
|
| 351 |
+
"learning_rate": 4.972536455871283e-05,
|
| 352 |
+
"loss": 0.5502,
|
| 353 |
+
"num_tokens": 164374316.0,
|
| 354 |
+
"step": 215
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.30013642564802184,
|
| 358 |
+
"grad_norm": 0.34186119793158315,
|
| 359 |
+
"learning_rate": 4.9698383087606136e-05,
|
| 360 |
+
"loss": 0.5296,
|
| 361 |
+
"num_tokens": 168223299.0,
|
| 362 |
+
"step": 220
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.3069577080491132,
|
| 366 |
+
"grad_norm": 0.29401395847274725,
|
| 367 |
+
"learning_rate": 4.9670146505401374e-05,
|
| 368 |
+
"loss": 0.5422,
|
| 369 |
+
"num_tokens": 172132000.0,
|
| 370 |
+
"step": 225
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 0.31377899045020463,
|
| 374 |
+
"grad_norm": 0.3668487339227577,
|
| 375 |
+
"learning_rate": 4.964065640861337e-05,
|
| 376 |
+
"loss": 0.5378,
|
| 377 |
+
"num_tokens": 176086331.0,
|
| 378 |
+
"step": 230
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 0.32060027285129605,
|
| 382 |
+
"grad_norm": 0.42051445325671166,
|
| 383 |
+
"learning_rate": 4.9609914464631466e-05,
|
| 384 |
+
"loss": 0.5301,
|
| 385 |
+
"num_tokens": 179917640.0,
|
| 386 |
+
"step": 235
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.3274215552523875,
|
| 390 |
+
"grad_norm": 0.3213672408738149,
|
| 391 |
+
"learning_rate": 4.957792241162531e-05,
|
| 392 |
+
"loss": 0.5427,
|
| 393 |
+
"num_tokens": 183746129.0,
|
| 394 |
+
"step": 240
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.33424283765347884,
|
| 398 |
+
"grad_norm": 0.40619564269997765,
|
| 399 |
+
"learning_rate": 4.954468205844648e-05,
|
| 400 |
+
"loss": 0.5514,
|
| 401 |
+
"num_tokens": 187699370.0,
|
| 402 |
+
"step": 245
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.34106412005457026,
|
| 406 |
+
"grad_norm": 0.37722270275447484,
|
| 407 |
+
"learning_rate": 4.95101952845263e-05,
|
| 408 |
+
"loss": 0.5336,
|
| 409 |
+
"num_tokens": 191512142.0,
|
| 410 |
+
"step": 250
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.3478854024556617,
|
| 414 |
+
"grad_norm": 0.3243135350430274,
|
| 415 |
+
"learning_rate": 4.94744640397695e-05,
|
| 416 |
+
"loss": 0.5347,
|
| 417 |
+
"num_tokens": 195367749.0,
|
| 418 |
+
"step": 255
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 0.35470668485675305,
|
| 422 |
+
"grad_norm": 0.3178884633626798,
|
| 423 |
+
"learning_rate": 4.943749034444406e-05,
|
| 424 |
+
"loss": 0.5235,
|
| 425 |
+
"num_tokens": 199215371.0,
|
| 426 |
+
"step": 260
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.3615279672578445,
|
| 430 |
+
"grad_norm": 0.34638109356483726,
|
| 431 |
+
"learning_rate": 4.9399276289066874e-05,
|
| 432 |
+
"loss": 0.543,
|
| 433 |
+
"num_tokens": 202903048.0,
|
| 434 |
+
"step": 265
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.3683492496589359,
|
| 438 |
+
"grad_norm": 0.3585730654530101,
|
| 439 |
+
"learning_rate": 4.935982403428563e-05,
|
| 440 |
+
"loss": 0.5318,
|
| 441 |
+
"num_tokens": 206761213.0,
|
| 442 |
+
"step": 270
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 0.37517053206002726,
|
| 446 |
+
"grad_norm": 0.3180094632527068,
|
| 447 |
+
"learning_rate": 4.931913581075661e-05,
|
| 448 |
+
"loss": 0.5278,
|
| 449 |
+
"num_tokens": 210585632.0,
|
| 450 |
+
"step": 275
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.3819918144611187,
|
| 454 |
+
"grad_norm": 0.30774207031129097,
|
| 455 |
+
"learning_rate": 4.92772139190186e-05,
|
| 456 |
+
"loss": 0.5337,
|
| 457 |
+
"num_tokens": 214261738.0,
|
| 458 |
+
"step": 280
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.3888130968622101,
|
| 462 |
+
"grad_norm": 0.30284100482119647,
|
| 463 |
+
"learning_rate": 4.923406072936276e-05,
|
| 464 |
+
"loss": 0.5382,
|
| 465 |
+
"num_tokens": 218144024.0,
|
| 466 |
+
"step": 285
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 0.3956343792633015,
|
| 470 |
+
"grad_norm": 0.35719024531210375,
|
| 471 |
+
"learning_rate": 4.918967868169866e-05,
|
| 472 |
+
"loss": 0.534,
|
| 473 |
+
"num_tokens": 222057295.0,
|
| 474 |
+
"step": 290
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.4024556616643929,
|
| 478 |
+
"grad_norm": 0.3454686804422533,
|
| 479 |
+
"learning_rate": 4.914407028541632e-05,
|
| 480 |
+
"loss": 0.5252,
|
| 481 |
+
"num_tokens": 225828573.0,
|
| 482 |
+
"step": 295
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.4092769440654843,
|
| 486 |
+
"grad_norm": 0.39488906682330294,
|
| 487 |
+
"learning_rate": 4.909723811924427e-05,
|
| 488 |
+
"loss": 0.525,
|
| 489 |
+
"num_tokens": 229710487.0,
|
| 490 |
+
"step": 300
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 0.41609822646657574,
|
| 494 |
+
"grad_norm": 0.33481820848360655,
|
| 495 |
+
"learning_rate": 4.9049184831103816e-05,
|
| 496 |
+
"loss": 0.5401,
|
| 497 |
+
"num_tokens": 233617525.0,
|
| 498 |
+
"step": 305
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 0.4229195088676671,
|
| 502 |
+
"grad_norm": 0.34678696659546243,
|
| 503 |
+
"learning_rate": 4.899991313795932e-05,
|
| 504 |
+
"loss": 0.5374,
|
| 505 |
+
"num_tokens": 237394471.0,
|
| 506 |
+
"step": 310
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.4297407912687585,
|
| 510 |
+
"grad_norm": 0.3511815797152881,
|
| 511 |
+
"learning_rate": 4.894942582566452e-05,
|
| 512 |
+
"loss": 0.5318,
|
| 513 |
+
"num_tokens": 241114261.0,
|
| 514 |
+
"step": 315
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.43656207366984995,
|
| 518 |
+
"grad_norm": 0.3283063276785904,
|
| 519 |
+
"learning_rate": 4.889772574880507e-05,
|
| 520 |
+
"loss": 0.5321,
|
| 521 |
+
"num_tokens": 244967987.0,
|
| 522 |
+
"step": 320
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.4433833560709413,
|
| 526 |
+
"grad_norm": 0.35201019136057504,
|
| 527 |
+
"learning_rate": 4.884481583053714e-05,
|
| 528 |
+
"loss": 0.5385,
|
| 529 |
+
"num_tokens": 248751095.0,
|
| 530 |
+
"step": 325
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.45020463847203274,
|
| 534 |
+
"grad_norm": 0.3885972425430605,
|
| 535 |
+
"learning_rate": 4.879069906242212e-05,
|
| 536 |
+
"loss": 0.526,
|
| 537 |
+
"num_tokens": 252607265.0,
|
| 538 |
+
"step": 330
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 0.45702592087312416,
|
| 542 |
+
"grad_norm": 0.3206664392088322,
|
| 543 |
+
"learning_rate": 4.8735378504257444e-05,
|
| 544 |
+
"loss": 0.5325,
|
| 545 |
+
"num_tokens": 256417909.0,
|
| 546 |
+
"step": 335
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 0.4638472032742155,
|
| 550 |
+
"grad_norm": 0.3744965651497676,
|
| 551 |
+
"learning_rate": 4.867885728390368e-05,
|
| 552 |
+
"loss": 0.5253,
|
| 553 |
+
"num_tokens": 260456429.0,
|
| 554 |
+
"step": 340
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.47066848567530695,
|
| 558 |
+
"grad_norm": 0.3189566677042572,
|
| 559 |
+
"learning_rate": 4.8621138597107574e-05,
|
| 560 |
+
"loss": 0.5185,
|
| 561 |
+
"num_tokens": 264287905.0,
|
| 562 |
+
"step": 345
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.47748976807639837,
|
| 566 |
+
"grad_norm": 0.36298464916985324,
|
| 567 |
+
"learning_rate": 4.856222570732144e-05,
|
| 568 |
+
"loss": 0.5272,
|
| 569 |
+
"num_tokens": 268098790.0,
|
| 570 |
+
"step": 350
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.4843110504774898,
|
| 574 |
+
"grad_norm": 0.34578336107449203,
|
| 575 |
+
"learning_rate": 4.8502121945518595e-05,
|
| 576 |
+
"loss": 0.5364,
|
| 577 |
+
"num_tokens": 271974065.0,
|
| 578 |
+
"step": 355
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.49113233287858116,
|
| 582 |
+
"grad_norm": 0.3774026933899186,
|
| 583 |
+
"learning_rate": 4.844083071000503e-05,
|
| 584 |
+
"loss": 0.5403,
|
| 585 |
+
"num_tokens": 275852045.0,
|
| 586 |
+
"step": 360
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 0.4979536152796726,
|
| 590 |
+
"grad_norm": 0.3620455108437782,
|
| 591 |
+
"learning_rate": 4.83783554662273e-05,
|
| 592 |
+
"loss": 0.5181,
|
| 593 |
+
"num_tokens": 279504484.0,
|
| 594 |
+
"step": 365
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.504774897680764,
|
| 598 |
+
"grad_norm": 0.3474757541708356,
|
| 599 |
+
"learning_rate": 4.831469974657655e-05,
|
| 600 |
+
"loss": 0.52,
|
| 601 |
+
"num_tokens": 283461546.0,
|
| 602 |
+
"step": 370
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.5115961800818554,
|
| 606 |
+
"grad_norm": 0.3584092026605748,
|
| 607 |
+
"learning_rate": 4.824986715018879e-05,
|
| 608 |
+
"loss": 0.5282,
|
| 609 |
+
"num_tokens": 287371918.0,
|
| 610 |
+
"step": 375
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 0.5184174624829468,
|
| 614 |
+
"grad_norm": 0.38542089863107043,
|
| 615 |
+
"learning_rate": 4.8183861342741424e-05,
|
| 616 |
+
"loss": 0.5144,
|
| 617 |
+
"num_tokens": 291057738.0,
|
| 618 |
+
"step": 380
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.5252387448840382,
|
| 622 |
+
"grad_norm": 0.33051941522122574,
|
| 623 |
+
"learning_rate": 4.811668605624598e-05,
|
| 624 |
+
"loss": 0.5238,
|
| 625 |
+
"num_tokens": 294881963.0,
|
| 626 |
+
"step": 385
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.5320600272851296,
|
| 630 |
+
"grad_norm": 0.327058733477332,
|
| 631 |
+
"learning_rate": 4.804834508883708e-05,
|
| 632 |
+
"loss": 0.5266,
|
| 633 |
+
"num_tokens": 298677895.0,
|
| 634 |
+
"step": 390
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.538881309686221,
|
| 638 |
+
"grad_norm": 0.28511337019961275,
|
| 639 |
+
"learning_rate": 4.7978842304557734e-05,
|
| 640 |
+
"loss": 0.5102,
|
| 641 |
+
"num_tokens": 302387985.0,
|
| 642 |
+
"step": 395
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 0.5457025920873124,
|
| 646 |
+
"grad_norm": 0.5721916623550405,
|
| 647 |
+
"learning_rate": 4.7908181633140815e-05,
|
| 648 |
+
"loss": 0.5328,
|
| 649 |
+
"num_tokens": 306130593.0,
|
| 650 |
+
"step": 400
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.5525238744884038,
|
| 654 |
+
"grad_norm": 0.40718220604246974,
|
| 655 |
+
"learning_rate": 4.78363670697869e-05,
|
| 656 |
+
"loss": 0.532,
|
| 657 |
+
"num_tokens": 309952008.0,
|
| 658 |
+
"step": 405
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.5593451568894953,
|
| 662 |
+
"grad_norm": 0.3804548334855215,
|
| 663 |
+
"learning_rate": 4.776340267493836e-05,
|
| 664 |
+
"loss": 0.5168,
|
| 665 |
+
"num_tokens": 313725489.0,
|
| 666 |
+
"step": 410
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 0.5661664392905866,
|
| 670 |
+
"grad_norm": 0.3286228983621768,
|
| 671 |
+
"learning_rate": 4.768929257404981e-05,
|
| 672 |
+
"loss": 0.5209,
|
| 673 |
+
"num_tokens": 317525148.0,
|
| 674 |
+
"step": 415
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.572987721691678,
|
| 678 |
+
"grad_norm": 0.33186583907144995,
|
| 679 |
+
"learning_rate": 4.761404095735482e-05,
|
| 680 |
+
"loss": 0.5208,
|
| 681 |
+
"num_tokens": 321410101.0,
|
| 682 |
+
"step": 420
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.5798090040927695,
|
| 686 |
+
"grad_norm": 0.3253640321282499,
|
| 687 |
+
"learning_rate": 4.753765207962902e-05,
|
| 688 |
+
"loss": 0.5158,
|
| 689 |
+
"num_tokens": 325102278.0,
|
| 690 |
+
"step": 425
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 0.5866302864938608,
|
| 694 |
+
"grad_norm": 0.3172673318697493,
|
| 695 |
+
"learning_rate": 4.7460130259949504e-05,
|
| 696 |
+
"loss": 0.5203,
|
| 697 |
+
"num_tokens": 328831071.0,
|
| 698 |
+
"step": 430
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.5934515688949522,
|
| 702 |
+
"grad_norm": 0.346434198786226,
|
| 703 |
+
"learning_rate": 4.738147988145068e-05,
|
| 704 |
+
"loss": 0.5099,
|
| 705 |
+
"num_tokens": 332767044.0,
|
| 706 |
+
"step": 435
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 0.6002728512960437,
|
| 710 |
+
"grad_norm": 0.3053374940441908,
|
| 711 |
+
"learning_rate": 4.730170539107636e-05,
|
| 712 |
+
"loss": 0.5073,
|
| 713 |
+
"num_tokens": 336595638.0,
|
| 714 |
+
"step": 440
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.607094133697135,
|
| 718 |
+
"grad_norm": 0.26630039820604245,
|
| 719 |
+
"learning_rate": 4.722081129932843e-05,
|
| 720 |
+
"loss": 0.5202,
|
| 721 |
+
"num_tokens": 340358925.0,
|
| 722 |
+
"step": 445
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.6139154160982264,
|
| 726 |
+
"grad_norm": 0.3310472023908457,
|
| 727 |
+
"learning_rate": 4.713880218001173e-05,
|
| 728 |
+
"loss": 0.519,
|
| 729 |
+
"num_tokens": 344239058.0,
|
| 730 |
+
"step": 450
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.6207366984993179,
|
| 734 |
+
"grad_norm": 0.2986769666429481,
|
| 735 |
+
"learning_rate": 4.7055682669975504e-05,
|
| 736 |
+
"loss": 0.5226,
|
| 737 |
+
"num_tokens": 348080585.0,
|
| 738 |
+
"step": 455
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.6275579809004093,
|
| 742 |
+
"grad_norm": 0.3067465928371488,
|
| 743 |
+
"learning_rate": 4.6971457468851207e-05,
|
| 744 |
+
"loss": 0.5161,
|
| 745 |
+
"num_tokens": 351817695.0,
|
| 746 |
+
"step": 460
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.6343792633015006,
|
| 750 |
+
"grad_norm": 0.32832217072725584,
|
| 751 |
+
"learning_rate": 4.68861313387868e-05,
|
| 752 |
+
"loss": 0.5107,
|
| 753 |
+
"num_tokens": 355639183.0,
|
| 754 |
+
"step": 465
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 0.6412005457025921,
|
| 758 |
+
"grad_norm": 0.35314800148659464,
|
| 759 |
+
"learning_rate": 4.67997091041775e-05,
|
| 760 |
+
"loss": 0.5292,
|
| 761 |
+
"num_tokens": 359437581.0,
|
| 762 |
+
"step": 470
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 0.6480218281036835,
|
| 766 |
+
"grad_norm": 0.31410739932640475,
|
| 767 |
+
"learning_rate": 4.671219565139295e-05,
|
| 768 |
+
"loss": 0.5052,
|
| 769 |
+
"num_tokens": 363189983.0,
|
| 770 |
+
"step": 475
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.654843110504775,
|
| 774 |
+
"grad_norm": 0.27562513330463034,
|
| 775 |
+
"learning_rate": 4.662359592850101e-05,
|
| 776 |
+
"loss": 0.5076,
|
| 777 |
+
"num_tokens": 366863209.0,
|
| 778 |
+
"step": 480
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 0.6616643929058663,
|
| 782 |
+
"grad_norm": 0.3567810768605306,
|
| 783 |
+
"learning_rate": 4.653391494498797e-05,
|
| 784 |
+
"loss": 0.5218,
|
| 785 |
+
"num_tokens": 370725860.0,
|
| 786 |
+
"step": 485
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.6684856753069577,
|
| 790 |
+
"grad_norm": 0.30442090678195616,
|
| 791 |
+
"learning_rate": 4.6443157771475276e-05,
|
| 792 |
+
"loss": 0.5071,
|
| 793 |
+
"num_tokens": 374566515.0,
|
| 794 |
+
"step": 490
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.6753069577080492,
|
| 798 |
+
"grad_norm": 0.267987058971068,
|
| 799 |
+
"learning_rate": 4.635132953943289e-05,
|
| 800 |
+
"loss": 0.4999,
|
| 801 |
+
"num_tokens": 378330489.0,
|
| 802 |
+
"step": 495
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.6821282401091405,
|
| 806 |
+
"grad_norm": 0.32102286772078575,
|
| 807 |
+
"learning_rate": 4.6258435440889106e-05,
|
| 808 |
+
"loss": 0.5045,
|
| 809 |
+
"num_tokens": 382103468.0,
|
| 810 |
+
"step": 500
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 0.6889495225102319,
|
| 814 |
+
"grad_norm": 0.32797699248595724,
|
| 815 |
+
"learning_rate": 4.6164480728137e-05,
|
| 816 |
+
"loss": 0.5128,
|
| 817 |
+
"num_tokens": 385837297.0,
|
| 818 |
+
"step": 505
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.6957708049113234,
|
| 822 |
+
"grad_norm": 0.3739420181528244,
|
| 823 |
+
"learning_rate": 4.6069470713437496e-05,
|
| 824 |
+
"loss": 0.5101,
|
| 825 |
+
"num_tokens": 389586680.0,
|
| 826 |
+
"step": 510
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 0.7025920873124147,
|
| 830 |
+
"grad_norm": 0.3177261550100274,
|
| 831 |
+
"learning_rate": 4.597341076871895e-05,
|
| 832 |
+
"loss": 0.5146,
|
| 833 |
+
"num_tokens": 393428760.0,
|
| 834 |
+
"step": 515
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.7094133697135061,
|
| 838 |
+
"grad_norm": 0.3258163855200761,
|
| 839 |
+
"learning_rate": 4.5876306325273475e-05,
|
| 840 |
+
"loss": 0.5263,
|
| 841 |
+
"num_tokens": 397158700.0,
|
| 842 |
+
"step": 520
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.7162346521145976,
|
| 846 |
+
"grad_norm": 0.3446090385934796,
|
| 847 |
+
"learning_rate": 4.577816287344981e-05,
|
| 848 |
+
"loss": 0.5107,
|
| 849 |
+
"num_tokens": 400923938.0,
|
| 850 |
+
"step": 525
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.723055934515689,
|
| 854 |
+
"grad_norm": 0.306719720813862,
|
| 855 |
+
"learning_rate": 4.5678985962342896e-05,
|
| 856 |
+
"loss": 0.5045,
|
| 857 |
+
"num_tokens": 404685655.0,
|
| 858 |
+
"step": 530
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 0.7298772169167803,
|
| 862 |
+
"grad_norm": 0.2921759737263205,
|
| 863 |
+
"learning_rate": 4.557878119948016e-05,
|
| 864 |
+
"loss": 0.5104,
|
| 865 |
+
"num_tokens": 408582799.0,
|
| 866 |
+
"step": 535
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.7366984993178718,
|
| 870 |
+
"grad_norm": 0.3130989626303816,
|
| 871 |
+
"learning_rate": 4.547755425050443e-05,
|
| 872 |
+
"loss": 0.5157,
|
| 873 |
+
"num_tokens": 412386009.0,
|
| 874 |
+
"step": 540
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 0.7435197817189632,
|
| 878 |
+
"grad_norm": 0.31463780976954164,
|
| 879 |
+
"learning_rate": 4.537531083885361e-05,
|
| 880 |
+
"loss": 0.5189,
|
| 881 |
+
"num_tokens": 416372039.0,
|
| 882 |
+
"step": 545
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 0.7503410641200545,
|
| 886 |
+
"grad_norm": 0.34415101422603883,
|
| 887 |
+
"learning_rate": 4.527205674543706e-05,
|
| 888 |
+
"loss": 0.5033,
|
| 889 |
+
"num_tokens": 420051975.0,
|
| 890 |
+
"step": 550
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.757162346521146,
|
| 894 |
+
"grad_norm": 0.3362194283915575,
|
| 895 |
+
"learning_rate": 4.516779780830876e-05,
|
| 896 |
+
"loss": 0.5038,
|
| 897 |
+
"num_tokens": 423685709.0,
|
| 898 |
+
"step": 555
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.7639836289222374,
|
| 902 |
+
"grad_norm": 0.2818223313948634,
|
| 903 |
+
"learning_rate": 4.5062539922337196e-05,
|
| 904 |
+
"loss": 0.4998,
|
| 905 |
+
"num_tokens": 427405904.0,
|
| 906 |
+
"step": 560
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.7708049113233287,
|
| 910 |
+
"grad_norm": 0.30589418056355644,
|
| 911 |
+
"learning_rate": 4.495628903887208e-05,
|
| 912 |
+
"loss": 0.5049,
|
| 913 |
+
"num_tokens": 431288378.0,
|
| 914 |
+
"step": 565
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.7776261937244202,
|
| 918 |
+
"grad_norm": 0.296990090107707,
|
| 919 |
+
"learning_rate": 4.484905116540786e-05,
|
| 920 |
+
"loss": 0.503,
|
| 921 |
+
"num_tokens": 435077995.0,
|
| 922 |
+
"step": 570
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.7844474761255116,
|
| 926 |
+
"grad_norm": 0.2985368221277938,
|
| 927 |
+
"learning_rate": 4.474083236524402e-05,
|
| 928 |
+
"loss": 0.5027,
|
| 929 |
+
"num_tokens": 439006864.0,
|
| 930 |
+
"step": 575
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 0.791268758526603,
|
| 934 |
+
"grad_norm": 0.3205820902932508,
|
| 935 |
+
"learning_rate": 4.463163875714234e-05,
|
| 936 |
+
"loss": 0.5178,
|
| 937 |
+
"num_tokens": 443045688.0,
|
| 938 |
+
"step": 580
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.7980900409276944,
|
| 942 |
+
"grad_norm": 0.3047870711403745,
|
| 943 |
+
"learning_rate": 4.452147651498079e-05,
|
| 944 |
+
"loss": 0.505,
|
| 945 |
+
"num_tokens": 446852018.0,
|
| 946 |
+
"step": 585
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 0.8049113233287858,
|
| 950 |
+
"grad_norm": 0.27987455949076323,
|
| 951 |
+
"learning_rate": 4.441035186740463e-05,
|
| 952 |
+
"loss": 0.5188,
|
| 953 |
+
"num_tokens": 450687287.0,
|
| 954 |
+
"step": 590
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.8117326057298773,
|
| 958 |
+
"grad_norm": 0.2879593820648548,
|
| 959 |
+
"learning_rate": 4.4298271097474075e-05,
|
| 960 |
+
"loss": 0.5043,
|
| 961 |
+
"num_tokens": 454483896.0,
|
| 962 |
+
"step": 595
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.8185538881309686,
|
| 966 |
+
"grad_norm": 0.27577549769516463,
|
| 967 |
+
"learning_rate": 4.4185240542309186e-05,
|
| 968 |
+
"loss": 0.5022,
|
| 969 |
+
"num_tokens": 458325861.0,
|
| 970 |
+
"step": 600
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 0.82537517053206,
|
| 974 |
+
"grad_norm": 0.2997594432337163,
|
| 975 |
+
"learning_rate": 4.407126659273148e-05,
|
| 976 |
+
"loss": 0.5076,
|
| 977 |
+
"num_tokens": 462055868.0,
|
| 978 |
+
"step": 605
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 0.8321964529331515,
|
| 982 |
+
"grad_norm": 0.26340007622161876,
|
| 983 |
+
"learning_rate": 4.395635569290259e-05,
|
| 984 |
+
"loss": 0.5017,
|
| 985 |
+
"num_tokens": 465973876.0,
|
| 986 |
+
"step": 610
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.8390177353342428,
|
| 990 |
+
"grad_norm": 0.26300808886198834,
|
| 991 |
+
"learning_rate": 4.384051433995996e-05,
|
| 992 |
+
"loss": 0.499,
|
| 993 |
+
"num_tokens": 469791498.0,
|
| 994 |
+
"step": 615
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 0.8458390177353342,
|
| 998 |
+
"grad_norm": 0.29111616742736374,
|
| 999 |
+
"learning_rate": 4.372374908364945e-05,
|
| 1000 |
+
"loss": 0.4965,
|
| 1001 |
+
"num_tokens": 473570581.0,
|
| 1002 |
+
"step": 620
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 0.8526603001364257,
|
| 1006 |
+
"grad_norm": 0.2786252827657706,
|
| 1007 |
+
"learning_rate": 4.3606066525954994e-05,
|
| 1008 |
+
"loss": 0.5125,
|
| 1009 |
+
"num_tokens": 477401353.0,
|
| 1010 |
+
"step": 625
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.859481582537517,
|
| 1014 |
+
"grad_norm": 0.2939906591983235,
|
| 1015 |
+
"learning_rate": 4.34874733207254e-05,
|
| 1016 |
+
"loss": 0.5116,
|
| 1017 |
+
"num_tokens": 481348892.0,
|
| 1018 |
+
"step": 630
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.8663028649386084,
|
| 1022 |
+
"grad_norm": 0.2881427084886996,
|
| 1023 |
+
"learning_rate": 4.336797617329806e-05,
|
| 1024 |
+
"loss": 0.5094,
|
| 1025 |
+
"num_tokens": 485171910.0,
|
| 1026 |
+
"step": 635
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 0.8731241473396999,
|
| 1030 |
+
"grad_norm": 0.2553738974972728,
|
| 1031 |
+
"learning_rate": 4.3247581840119814e-05,
|
| 1032 |
+
"loss": 0.4943,
|
| 1033 |
+
"num_tokens": 489059548.0,
|
| 1034 |
+
"step": 640
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.8799454297407913,
|
| 1038 |
+
"grad_norm": 0.2838429765846332,
|
| 1039 |
+
"learning_rate": 4.312629712836502e-05,
|
| 1040 |
+
"loss": 0.4975,
|
| 1041 |
+
"num_tokens": 492893029.0,
|
| 1042 |
+
"step": 645
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 0.8867667121418826,
|
| 1046 |
+
"grad_norm": 0.301109628714455,
|
| 1047 |
+
"learning_rate": 4.3004128895550595e-05,
|
| 1048 |
+
"loss": 0.505,
|
| 1049 |
+
"num_tokens": 496905674.0,
|
| 1050 |
+
"step": 650
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 0.8935879945429741,
|
| 1054 |
+
"grad_norm": 0.2861782459282683,
|
| 1055 |
+
"learning_rate": 4.288108404914833e-05,
|
| 1056 |
+
"loss": 0.4999,
|
| 1057 |
+
"num_tokens": 500864542.0,
|
| 1058 |
+
"step": 655
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.9004092769440655,
|
| 1062 |
+
"grad_norm": 0.339733593259804,
|
| 1063 |
+
"learning_rate": 4.2757169546194315e-05,
|
| 1064 |
+
"loss": 0.5037,
|
| 1065 |
+
"num_tokens": 504810366.0,
|
| 1066 |
+
"step": 660
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.9072305593451568,
|
| 1070 |
+
"grad_norm": 0.3473691590981822,
|
| 1071 |
+
"learning_rate": 4.263239239289559e-05,
|
| 1072 |
+
"loss": 0.4998,
|
| 1073 |
+
"num_tokens": 508607232.0,
|
| 1074 |
+
"step": 665
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.9140518417462483,
|
| 1078 |
+
"grad_norm": 0.3171779108710887,
|
| 1079 |
+
"learning_rate": 4.2506759644233994e-05,
|
| 1080 |
+
"loss": 0.5067,
|
| 1081 |
+
"num_tokens": 512474084.0,
|
| 1082 |
+
"step": 670
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.9208731241473397,
|
| 1086 |
+
"grad_norm": 0.28330202092522305,
|
| 1087 |
+
"learning_rate": 4.238027840356733e-05,
|
| 1088 |
+
"loss": 0.4972,
|
| 1089 |
+
"num_tokens": 516216104.0,
|
| 1090 |
+
"step": 675
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 0.927694406548431,
|
| 1094 |
+
"grad_norm": 0.27582823397924155,
|
| 1095 |
+
"learning_rate": 4.2252955822227656e-05,
|
| 1096 |
+
"loss": 0.5065,
|
| 1097 |
+
"num_tokens": 519966345.0,
|
| 1098 |
+
"step": 680
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.9345156889495225,
|
| 1102 |
+
"grad_norm": 0.2768074983589449,
|
| 1103 |
+
"learning_rate": 4.212479909911699e-05,
|
| 1104 |
+
"loss": 0.4921,
|
| 1105 |
+
"num_tokens": 523793837.0,
|
| 1106 |
+
"step": 685
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.9413369713506139,
|
| 1110 |
+
"grad_norm": 0.26048610160343333,
|
| 1111 |
+
"learning_rate": 4.199581548030028e-05,
|
| 1112 |
+
"loss": 0.5055,
|
| 1113 |
+
"num_tokens": 527666864.0,
|
| 1114 |
+
"step": 690
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 0.9481582537517054,
|
| 1118 |
+
"grad_norm": 0.2999705117009638,
|
| 1119 |
+
"learning_rate": 4.186601225859567e-05,
|
| 1120 |
+
"loss": 0.5028,
|
| 1121 |
+
"num_tokens": 531466359.0,
|
| 1122 |
+
"step": 695
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.9549795361527967,
|
| 1126 |
+
"grad_norm": 0.3022254979765851,
|
| 1127 |
+
"learning_rate": 4.1735396773162244e-05,
|
| 1128 |
+
"loss": 0.5052,
|
| 1129 |
+
"num_tokens": 535253217.0,
|
| 1130 |
+
"step": 700
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.9618008185538881,
|
| 1134 |
+
"grad_norm": 0.29139393827615595,
|
| 1135 |
+
"learning_rate": 4.1603976409084954e-05,
|
| 1136 |
+
"loss": 0.4891,
|
| 1137 |
+
"num_tokens": 539137436.0,
|
| 1138 |
+
"step": 705
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 0.9686221009549796,
|
| 1142 |
+
"grad_norm": 0.28098849806023163,
|
| 1143 |
+
"learning_rate": 4.1471758596957146e-05,
|
| 1144 |
+
"loss": 0.4995,
|
| 1145 |
+
"num_tokens": 542824098.0,
|
| 1146 |
+
"step": 710
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 0.975443383356071,
|
| 1150 |
+
"grad_norm": 0.2894956414917047,
|
| 1151 |
+
"learning_rate": 4.1338750812460384e-05,
|
| 1152 |
+
"loss": 0.5105,
|
| 1153 |
+
"num_tokens": 546578047.0,
|
| 1154 |
+
"step": 715
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.9822646657571623,
|
| 1158 |
+
"grad_norm": 0.26624260116360937,
|
| 1159 |
+
"learning_rate": 4.1204960575941794e-05,
|
| 1160 |
+
"loss": 0.4961,
|
| 1161 |
+
"num_tokens": 550377811.0,
|
| 1162 |
+
"step": 720
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 0.9890859481582538,
|
| 1166 |
+
"grad_norm": 0.29134798209278545,
|
| 1167 |
+
"learning_rate": 4.1070395451988855e-05,
|
| 1168 |
+
"loss": 0.4926,
|
| 1169 |
+
"num_tokens": 554224024.0,
|
| 1170 |
+
"step": 725
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 0.9959072305593452,
|
| 1174 |
+
"grad_norm": 0.31317560859597543,
|
| 1175 |
+
"learning_rate": 4.093506304900168e-05,
|
| 1176 |
+
"loss": 0.5002,
|
| 1177 |
+
"num_tokens": 558001366.0,
|
| 1178 |
+
"step": 730
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 1.0027285129604366,
|
| 1182 |
+
"grad_norm": 0.3379410163089462,
|
| 1183 |
+
"learning_rate": 4.079897101876286e-05,
|
| 1184 |
+
"loss": 0.4875,
|
| 1185 |
+
"num_tokens": 561842493.0,
|
| 1186 |
+
"step": 735
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.009549795361528,
|
| 1190 |
+
"grad_norm": 0.2931355067947464,
|
| 1191 |
+
"learning_rate": 4.066212705600477e-05,
|
| 1192 |
+
"loss": 0.4663,
|
| 1193 |
+
"num_tokens": 565569218.0,
|
| 1194 |
+
"step": 740
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 1.0163710777626194,
|
| 1198 |
+
"grad_norm": 0.3470937085195407,
|
| 1199 |
+
"learning_rate": 4.052453889797457e-05,
|
| 1200 |
+
"loss": 0.4747,
|
| 1201 |
+
"num_tokens": 569401220.0,
|
| 1202 |
+
"step": 745
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 1.0231923601637107,
|
| 1206 |
+
"grad_norm": 0.3558247142601529,
|
| 1207 |
+
"learning_rate": 4.038621432399671e-05,
|
| 1208 |
+
"loss": 0.475,
|
| 1209 |
+
"num_tokens": 573382420.0,
|
| 1210 |
+
"step": 750
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 1.030013642564802,
|
| 1214 |
+
"grad_norm": 0.28310978477738025,
|
| 1215 |
+
"learning_rate": 4.024716115503307e-05,
|
| 1216 |
+
"loss": 0.4653,
|
| 1217 |
+
"num_tokens": 577211964.0,
|
| 1218 |
+
"step": 755
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 1.0368349249658937,
|
| 1222 |
+
"grad_norm": 0.2777221889564963,
|
| 1223 |
+
"learning_rate": 4.010738725324075e-05,
|
| 1224 |
+
"loss": 0.4671,
|
| 1225 |
+
"num_tokens": 581185526.0,
|
| 1226 |
+
"step": 760
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.043656207366985,
|
| 1230 |
+
"grad_norm": 0.2884904481361175,
|
| 1231 |
+
"learning_rate": 3.9966900521527596e-05,
|
| 1232 |
+
"loss": 0.4692,
|
| 1233 |
+
"num_tokens": 584961788.0,
|
| 1234 |
+
"step": 765
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 1.0504774897680764,
|
| 1238 |
+
"grad_norm": 0.2989739733410286,
|
| 1239 |
+
"learning_rate": 3.9825708903105284e-05,
|
| 1240 |
+
"loss": 0.4779,
|
| 1241 |
+
"num_tokens": 588808530.0,
|
| 1242 |
+
"step": 770
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.0572987721691678,
|
| 1246 |
+
"grad_norm": 0.3229871452710858,
|
| 1247 |
+
"learning_rate": 3.968382038104027e-05,
|
| 1248 |
+
"loss": 0.4582,
|
| 1249 |
+
"num_tokens": 592600556.0,
|
| 1250 |
+
"step": 775
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 1.0641200545702592,
|
| 1254 |
+
"grad_norm": 0.2795338016441354,
|
| 1255 |
+
"learning_rate": 3.9541242977802395e-05,
|
| 1256 |
+
"loss": 0.4656,
|
| 1257 |
+
"num_tokens": 596480393.0,
|
| 1258 |
+
"step": 780
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 1.0709413369713505,
|
| 1262 |
+
"grad_norm": 0.28782990015102894,
|
| 1263 |
+
"learning_rate": 3.939798475481128e-05,
|
| 1264 |
+
"loss": 0.4699,
|
| 1265 |
+
"num_tokens": 600450795.0,
|
| 1266 |
+
"step": 785
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 1.077762619372442,
|
| 1270 |
+
"grad_norm": 0.2666268848706678,
|
| 1271 |
+
"learning_rate": 3.925405381198055e-05,
|
| 1272 |
+
"loss": 0.472,
|
| 1273 |
+
"num_tokens": 604180876.0,
|
| 1274 |
+
"step": 790
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.0845839017735335,
|
| 1278 |
+
"grad_norm": 0.2996103352872881,
|
| 1279 |
+
"learning_rate": 3.9109458287259834e-05,
|
| 1280 |
+
"loss": 0.4652,
|
| 1281 |
+
"num_tokens": 607934610.0,
|
| 1282 |
+
"step": 795
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.0914051841746248,
|
| 1286 |
+
"grad_norm": 0.27968650209717966,
|
| 1287 |
+
"learning_rate": 3.896420635617469e-05,
|
| 1288 |
+
"loss": 0.4676,
|
| 1289 |
+
"num_tokens": 611651241.0,
|
| 1290 |
+
"step": 800
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 1.0982264665757162,
|
| 1294 |
+
"grad_norm": 0.3273962994074201,
|
| 1295 |
+
"learning_rate": 3.881830623136429e-05,
|
| 1296 |
+
"loss": 0.4655,
|
| 1297 |
+
"num_tokens": 615434308.0,
|
| 1298 |
+
"step": 805
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.1050477489768076,
|
| 1302 |
+
"grad_norm": 0.26789689269716943,
|
| 1303 |
+
"learning_rate": 3.867176616211711e-05,
|
| 1304 |
+
"loss": 0.4637,
|
| 1305 |
+
"num_tokens": 619368833.0,
|
| 1306 |
+
"step": 810
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 1.111869031377899,
|
| 1310 |
+
"grad_norm": 0.27962594173651023,
|
| 1311 |
+
"learning_rate": 3.8524594433904524e-05,
|
| 1312 |
+
"loss": 0.4618,
|
| 1313 |
+
"num_tokens": 623126366.0,
|
| 1314 |
+
"step": 815
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 1.1186903137789905,
|
| 1318 |
+
"grad_norm": 0.26130675702636497,
|
| 1319 |
+
"learning_rate": 3.837679936791227e-05,
|
| 1320 |
+
"loss": 0.4744,
|
| 1321 |
+
"num_tokens": 627107529.0,
|
| 1322 |
+
"step": 820
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.125511596180082,
|
| 1326 |
+
"grad_norm": 0.255697817940833,
|
| 1327 |
+
"learning_rate": 3.822838932057006e-05,
|
| 1328 |
+
"loss": 0.4627,
|
| 1329 |
+
"num_tokens": 630903442.0,
|
| 1330 |
+
"step": 825
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 1.1323328785811733,
|
| 1334 |
+
"grad_norm": 0.28299476665523327,
|
| 1335 |
+
"learning_rate": 3.807937268307902e-05,
|
| 1336 |
+
"loss": 0.4585,
|
| 1337 |
+
"num_tokens": 634517705.0,
|
| 1338 |
+
"step": 830
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.1391541609822646,
|
| 1342 |
+
"grad_norm": 0.2869277101946544,
|
| 1343 |
+
"learning_rate": 3.792975788093732e-05,
|
| 1344 |
+
"loss": 0.4666,
|
| 1345 |
+
"num_tokens": 638096091.0,
|
| 1346 |
+
"step": 835
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.145975443383356,
|
| 1350 |
+
"grad_norm": 0.29785970150251395,
|
| 1351 |
+
"learning_rate": 3.77795533734637e-05,
|
| 1352 |
+
"loss": 0.4662,
|
| 1353 |
+
"num_tokens": 642002612.0,
|
| 1354 |
+
"step": 840
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.1527967257844476,
|
| 1358 |
+
"grad_norm": 0.2739981453811102,
|
| 1359 |
+
"learning_rate": 3.762876765331927e-05,
|
| 1360 |
+
"loss": 0.4716,
|
| 1361 |
+
"num_tokens": 645977691.0,
|
| 1362 |
+
"step": 845
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 1.159618008185539,
|
| 1366 |
+
"grad_norm": 0.2863210693637654,
|
| 1367 |
+
"learning_rate": 3.7477409246027244e-05,
|
| 1368 |
+
"loss": 0.4647,
|
| 1369 |
+
"num_tokens": 649827816.0,
|
| 1370 |
+
"step": 850
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 1.1664392905866303,
|
| 1374 |
+
"grad_norm": 0.3057072927224209,
|
| 1375 |
+
"learning_rate": 3.7325486709490974e-05,
|
| 1376 |
+
"loss": 0.4742,
|
| 1377 |
+
"num_tokens": 653510445.0,
|
| 1378 |
+
"step": 855
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 1.1732605729877217,
|
| 1382 |
+
"grad_norm": 0.2850477095012747,
|
| 1383 |
+
"learning_rate": 3.717300863351004e-05,
|
| 1384 |
+
"loss": 0.469,
|
| 1385 |
+
"num_tokens": 657384998.0,
|
| 1386 |
+
"step": 860
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 1.180081855388813,
|
| 1390 |
+
"grad_norm": 0.2834741618680964,
|
| 1391 |
+
"learning_rate": 3.701998363929458e-05,
|
| 1392 |
+
"loss": 0.4644,
|
| 1393 |
+
"num_tokens": 661293472.0,
|
| 1394 |
+
"step": 865
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.1869031377899044,
|
| 1398 |
+
"grad_norm": 0.29014683601132446,
|
| 1399 |
+
"learning_rate": 3.686642037897784e-05,
|
| 1400 |
+
"loss": 0.4726,
|
| 1401 |
+
"num_tokens": 665091673.0,
|
| 1402 |
+
"step": 870
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.1937244201909958,
|
| 1406 |
+
"grad_norm": 0.2748758522694655,
|
| 1407 |
+
"learning_rate": 3.671232753512701e-05,
|
| 1408 |
+
"loss": 0.4705,
|
| 1409 |
+
"num_tokens": 668922968.0,
|
| 1410 |
+
"step": 875
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.2005457025920874,
|
| 1414 |
+
"grad_norm": 0.24796018921153712,
|
| 1415 |
+
"learning_rate": 3.6557713820252245e-05,
|
| 1416 |
+
"loss": 0.462,
|
| 1417 |
+
"num_tokens": 672846634.0,
|
| 1418 |
+
"step": 880
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 1.2073669849931787,
|
| 1422 |
+
"grad_norm": 0.30805411714650005,
|
| 1423 |
+
"learning_rate": 3.640258797631411e-05,
|
| 1424 |
+
"loss": 0.4604,
|
| 1425 |
+
"num_tokens": 676655826.0,
|
| 1426 |
+
"step": 885
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 1.21418826739427,
|
| 1430 |
+
"grad_norm": 0.29455075507488565,
|
| 1431 |
+
"learning_rate": 3.624695877422928e-05,
|
| 1432 |
+
"loss": 0.4766,
|
| 1433 |
+
"num_tokens": 680322801.0,
|
| 1434 |
+
"step": 890
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 1.2210095497953615,
|
| 1438 |
+
"grad_norm": 8.814550543938628,
|
| 1439 |
+
"learning_rate": 3.6090835013374616e-05,
|
| 1440 |
+
"loss": 0.4715,
|
| 1441 |
+
"num_tokens": 684119272.0,
|
| 1442 |
+
"step": 895
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.2278308321964528,
|
| 1446 |
+
"grad_norm": 0.871271666835307,
|
| 1447 |
+
"learning_rate": 3.5934225521089696e-05,
|
| 1448 |
+
"loss": 0.4652,
|
| 1449 |
+
"num_tokens": 687991406.0,
|
| 1450 |
+
"step": 900
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.2346521145975444,
|
| 1454 |
+
"grad_norm": 0.258382830995481,
|
| 1455 |
+
"learning_rate": 3.57771391521776e-05,
|
| 1456 |
+
"loss": 0.4747,
|
| 1457 |
+
"num_tokens": 691695054.0,
|
| 1458 |
+
"step": 905
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.2414733969986358,
|
| 1462 |
+
"grad_norm": 0.25064252265781106,
|
| 1463 |
+
"learning_rate": 3.56195847884044e-05,
|
| 1464 |
+
"loss": 0.4657,
|
| 1465 |
+
"num_tokens": 695621817.0,
|
| 1466 |
+
"step": 910
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.2482946793997272,
|
| 1470 |
+
"grad_norm": 0.26182463275561424,
|
| 1471 |
+
"learning_rate": 3.546157133799684e-05,
|
| 1472 |
+
"loss": 0.4535,
|
| 1473 |
+
"num_tokens": 699480008.0,
|
| 1474 |
+
"step": 915
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 1.2551159618008185,
|
| 1478 |
+
"grad_norm": 0.2873108568050052,
|
| 1479 |
+
"learning_rate": 3.5303107735138773e-05,
|
| 1480 |
+
"loss": 0.448,
|
| 1481 |
+
"num_tokens": 703195130.0,
|
| 1482 |
+
"step": 920
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 1.26193724420191,
|
| 1486 |
+
"grad_norm": 0.28664281506831896,
|
| 1487 |
+
"learning_rate": 3.5144202939465926e-05,
|
| 1488 |
+
"loss": 0.4664,
|
| 1489 |
+
"num_tokens": 707113602.0,
|
| 1490 |
+
"step": 925
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.2687585266030013,
|
| 1494 |
+
"grad_norm": 0.28926863378235584,
|
| 1495 |
+
"learning_rate": 3.4984865935559385e-05,
|
| 1496 |
+
"loss": 0.4636,
|
| 1497 |
+
"num_tokens": 710925367.0,
|
| 1498 |
+
"step": 930
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.2755798090040928,
|
| 1502 |
+
"grad_norm": 0.2821085120527458,
|
| 1503 |
+
"learning_rate": 3.482510573243756e-05,
|
| 1504 |
+
"loss": 0.4618,
|
| 1505 |
+
"num_tokens": 714722146.0,
|
| 1506 |
+
"step": 935
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 1.2824010914051842,
|
| 1510 |
+
"grad_norm": 0.26610575452010354,
|
| 1511 |
+
"learning_rate": 3.466493136304682e-05,
|
| 1512 |
+
"loss": 0.4585,
|
| 1513 |
+
"num_tokens": 718555460.0,
|
| 1514 |
+
"step": 940
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.2892223738062756,
|
| 1518 |
+
"grad_norm": 0.2835087759081065,
|
| 1519 |
+
"learning_rate": 3.450435188375077e-05,
|
| 1520 |
+
"loss": 0.4614,
|
| 1521 |
+
"num_tokens": 722379411.0,
|
| 1522 |
+
"step": 945
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.296043656207367,
|
| 1526 |
+
"grad_norm": 0.26785185762300534,
|
| 1527 |
+
"learning_rate": 3.4343376373818186e-05,
|
| 1528 |
+
"loss": 0.471,
|
| 1529 |
+
"num_tokens": 726285777.0,
|
| 1530 |
+
"step": 950
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 1.3028649386084583,
|
| 1534 |
+
"grad_norm": 0.265918636255517,
|
| 1535 |
+
"learning_rate": 3.418201393490969e-05,
|
| 1536 |
+
"loss": 0.4762,
|
| 1537 |
+
"num_tokens": 730064545.0,
|
| 1538 |
+
"step": 955
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 1.30968622100955,
|
| 1542 |
+
"grad_norm": 0.2622203591646629,
|
| 1543 |
+
"learning_rate": 3.4020273690563115e-05,
|
| 1544 |
+
"loss": 0.4597,
|
| 1545 |
+
"num_tokens": 733841294.0,
|
| 1546 |
+
"step": 960
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 1.3165075034106413,
|
| 1550 |
+
"grad_norm": 0.2727075473902522,
|
| 1551 |
+
"learning_rate": 3.385816478567765e-05,
|
| 1552 |
+
"loss": 0.4709,
|
| 1553 |
+
"num_tokens": 737760285.0,
|
| 1554 |
+
"step": 965
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 1.3233287858117326,
|
| 1558 |
+
"grad_norm": 0.30765863393818926,
|
| 1559 |
+
"learning_rate": 3.36956963859968e-05,
|
| 1560 |
+
"loss": 0.4705,
|
| 1561 |
+
"num_tokens": 741596226.0,
|
| 1562 |
+
"step": 970
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 1.330150068212824,
|
| 1566 |
+
"grad_norm": 0.3065429275697597,
|
| 1567 |
+
"learning_rate": 3.3532877677590155e-05,
|
| 1568 |
+
"loss": 0.4626,
|
| 1569 |
+
"num_tokens": 745437763.0,
|
| 1570 |
+
"step": 975
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.3369713506139154,
|
| 1574 |
+
"grad_norm": 0.2957611738142281,
|
| 1575 |
+
"learning_rate": 3.336971786633396e-05,
|
| 1576 |
+
"loss": 0.4623,
|
| 1577 |
+
"num_tokens": 749302808.0,
|
| 1578 |
+
"step": 980
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.3437926330150067,
|
| 1582 |
+
"grad_norm": 0.2955717204124572,
|
| 1583 |
+
"learning_rate": 3.3206226177390666e-05,
|
| 1584 |
+
"loss": 0.4665,
|
| 1585 |
+
"num_tokens": 753127588.0,
|
| 1586 |
+
"step": 985
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.350613915416098,
|
| 1590 |
+
"grad_norm": 0.2628487978573347,
|
| 1591 |
+
"learning_rate": 3.3042411854687276e-05,
|
| 1592 |
+
"loss": 0.4663,
|
| 1593 |
+
"num_tokens": 756971687.0,
|
| 1594 |
+
"step": 990
|
| 1595 |
+
},
|
| 1596 |
+
{
|
| 1597 |
+
"epoch": 1.3574351978171897,
|
| 1598 |
+
"grad_norm": 0.2457783122325923,
|
| 1599 |
+
"learning_rate": 3.287828416039276e-05,
|
| 1600 |
+
"loss": 0.4574,
|
| 1601 |
+
"num_tokens": 760697602.0,
|
| 1602 |
+
"step": 995
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 1.364256480218281,
|
| 1606 |
+
"grad_norm": 0.2488103415192936,
|
| 1607 |
+
"learning_rate": 3.271385237439427e-05,
|
| 1608 |
+
"loss": 0.4636,
|
| 1609 |
+
"num_tokens": 764490647.0,
|
| 1610 |
+
"step": 1000
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 1.3710777626193724,
|
| 1614 |
+
"grad_norm": 0.2661325953924744,
|
| 1615 |
+
"learning_rate": 3.254912579377256e-05,
|
| 1616 |
+
"loss": 0.4629,
|
| 1617 |
+
"num_tokens": 768330897.0,
|
| 1618 |
+
"step": 1005
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.3778990450204638,
|
| 1622 |
+
"grad_norm": 0.25618143079474864,
|
| 1623 |
+
"learning_rate": 3.23841137322762e-05,
|
| 1624 |
+
"loss": 0.464,
|
| 1625 |
+
"num_tokens": 772161013.0,
|
| 1626 |
+
"step": 1010
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.3847203274215554,
|
| 1630 |
+
"grad_norm": 0.2685440959203762,
|
| 1631 |
+
"learning_rate": 3.221882551979509e-05,
|
| 1632 |
+
"loss": 0.4709,
|
| 1633 |
+
"num_tokens": 776138853.0,
|
| 1634 |
+
"step": 1015
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.3915416098226467,
|
| 1638 |
+
"grad_norm": 0.23931689676797138,
|
| 1639 |
+
"learning_rate": 3.2053270501832853e-05,
|
| 1640 |
+
"loss": 0.4583,
|
| 1641 |
+
"num_tokens": 779984056.0,
|
| 1642 |
+
"step": 1020
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 1.398362892223738,
|
| 1646 |
+
"grad_norm": 0.266990380836569,
|
| 1647 |
+
"learning_rate": 3.18874580389785e-05,
|
| 1648 |
+
"loss": 0.4603,
|
| 1649 |
+
"num_tokens": 783898134.0,
|
| 1650 |
+
"step": 1025
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 1.4051841746248295,
|
| 1654 |
+
"grad_norm": 0.26893782452785026,
|
| 1655 |
+
"learning_rate": 3.172139750637711e-05,
|
| 1656 |
+
"loss": 0.462,
|
| 1657 |
+
"num_tokens": 787897711.0,
|
| 1658 |
+
"step": 1030
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 1.4120054570259208,
|
| 1662 |
+
"grad_norm": 0.2602735103785085,
|
| 1663 |
+
"learning_rate": 3.155509829319982e-05,
|
| 1664 |
+
"loss": 0.4678,
|
| 1665 |
+
"num_tokens": 791719537.0,
|
| 1666 |
+
"step": 1035
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 1.4188267394270122,
|
| 1670 |
+
"grad_norm": 0.29800812285353256,
|
| 1671 |
+
"learning_rate": 3.138856980211289e-05,
|
| 1672 |
+
"loss": 0.4626,
|
| 1673 |
+
"num_tokens": 795317290.0,
|
| 1674 |
+
"step": 1040
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 1.4256480218281036,
|
| 1678 |
+
"grad_norm": 0.26161467639459157,
|
| 1679 |
+
"learning_rate": 3.1221821448746126e-05,
|
| 1680 |
+
"loss": 0.4621,
|
| 1681 |
+
"num_tokens": 799156629.0,
|
| 1682 |
+
"step": 1045
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.4324693042291952,
|
| 1686 |
+
"grad_norm": 0.24989464458356864,
|
| 1687 |
+
"learning_rate": 3.105486266116048e-05,
|
| 1688 |
+
"loss": 0.4526,
|
| 1689 |
+
"num_tokens": 803069636.0,
|
| 1690 |
+
"step": 1050
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.4392905866302865,
|
| 1694 |
+
"grad_norm": 0.2802567216692825,
|
| 1695 |
+
"learning_rate": 3.088770287931501e-05,
|
| 1696 |
+
"loss": 0.4629,
|
| 1697 |
+
"num_tokens": 807046765.0,
|
| 1698 |
+
"step": 1055
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 1.446111869031378,
|
| 1702 |
+
"grad_norm": 0.24008936822242133,
|
| 1703 |
+
"learning_rate": 3.072035155453311e-05,
|
| 1704 |
+
"loss": 0.4616,
|
| 1705 |
+
"num_tokens": 810878242.0,
|
| 1706 |
+
"step": 1060
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 1.4529331514324693,
|
| 1710 |
+
"grad_norm": 0.2539318300904362,
|
| 1711 |
+
"learning_rate": 3.055281814896812e-05,
|
| 1712 |
+
"loss": 0.4612,
|
| 1713 |
+
"num_tokens": 814684093.0,
|
| 1714 |
+
"step": 1065
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 1.4597544338335606,
|
| 1718 |
+
"grad_norm": 0.2565319360647305,
|
| 1719 |
+
"learning_rate": 3.038511213506835e-05,
|
| 1720 |
+
"loss": 0.4587,
|
| 1721 |
+
"num_tokens": 818487158.0,
|
| 1722 |
+
"step": 1070
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.4665757162346522,
|
| 1726 |
+
"grad_norm": 0.2780910655918512,
|
| 1727 |
+
"learning_rate": 3.0217242995041544e-05,
|
| 1728 |
+
"loss": 0.4586,
|
| 1729 |
+
"num_tokens": 822359745.0,
|
| 1730 |
+
"step": 1075
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 1.4733969986357436,
|
| 1734 |
+
"grad_norm": 0.2407671631402225,
|
| 1735 |
+
"learning_rate": 3.004922022031866e-05,
|
| 1736 |
+
"loss": 0.4569,
|
| 1737 |
+
"num_tokens": 826031854.0,
|
| 1738 |
+
"step": 1080
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.480218281036835,
|
| 1742 |
+
"grad_norm": 0.264977891423622,
|
| 1743 |
+
"learning_rate": 2.988105331101726e-05,
|
| 1744 |
+
"loss": 0.4627,
|
| 1745 |
+
"num_tokens": 829831729.0,
|
| 1746 |
+
"step": 1085
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 1.4870395634379263,
|
| 1750 |
+
"grad_norm": 0.26279547832113415,
|
| 1751 |
+
"learning_rate": 2.9712751775404386e-05,
|
| 1752 |
+
"loss": 0.4608,
|
| 1753 |
+
"num_tokens": 833559495.0,
|
| 1754 |
+
"step": 1090
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 1.4938608458390177,
|
| 1758 |
+
"grad_norm": 0.26250576691518845,
|
| 1759 |
+
"learning_rate": 2.9544325129358925e-05,
|
| 1760 |
+
"loss": 0.4718,
|
| 1761 |
+
"num_tokens": 837452717.0,
|
| 1762 |
+
"step": 1095
|
| 1763 |
+
},
|
| 1764 |
+
{
|
| 1765 |
+
"epoch": 1.500682128240109,
|
| 1766 |
+
"grad_norm": 0.2564281846472605,
|
| 1767 |
+
"learning_rate": 2.937578289583362e-05,
|
| 1768 |
+
"loss": 0.4657,
|
| 1769 |
+
"num_tokens": 841494894.0,
|
| 1770 |
+
"step": 1100
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 1.5075034106412004,
|
| 1774 |
+
"grad_norm": 0.2754176412312844,
|
| 1775 |
+
"learning_rate": 2.920713460431658e-05,
|
| 1776 |
+
"loss": 0.4564,
|
| 1777 |
+
"num_tokens": 845244637.0,
|
| 1778 |
+
"step": 1105
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 1.514324693042292,
|
| 1782 |
+
"grad_norm": 0.24271490421490427,
|
| 1783 |
+
"learning_rate": 2.9038389790292497e-05,
|
| 1784 |
+
"loss": 0.4618,
|
| 1785 |
+
"num_tokens": 849040114.0,
|
| 1786 |
+
"step": 1110
|
| 1787 |
+
},
|
| 1788 |
+
{
|
| 1789 |
+
"epoch": 1.5211459754433834,
|
| 1790 |
+
"grad_norm": 0.26242407204309637,
|
| 1791 |
+
"learning_rate": 2.8869557994703527e-05,
|
| 1792 |
+
"loss": 0.4613,
|
| 1793 |
+
"num_tokens": 852855908.0,
|
| 1794 |
+
"step": 1115
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 1.5279672578444747,
|
| 1798 |
+
"grad_norm": 0.23892736672685205,
|
| 1799 |
+
"learning_rate": 2.8700648763409815e-05,
|
| 1800 |
+
"loss": 0.4605,
|
| 1801 |
+
"num_tokens": 856606710.0,
|
| 1802 |
+
"step": 1120
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 1.5347885402455663,
|
| 1806 |
+
"grad_norm": 0.2535034329296099,
|
| 1807 |
+
"learning_rate": 2.853167164664976e-05,
|
| 1808 |
+
"loss": 0.4691,
|
| 1809 |
+
"num_tokens": 860530719.0,
|
| 1810 |
+
"step": 1125
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 1.5416098226466577,
|
| 1814 |
+
"grad_norm": 0.26731315936513816,
|
| 1815 |
+
"learning_rate": 2.8362636198500064e-05,
|
| 1816 |
+
"loss": 0.4596,
|
| 1817 |
+
"num_tokens": 864403759.0,
|
| 1818 |
+
"step": 1130
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 1.548431105047749,
|
| 1822 |
+
"grad_norm": 0.23363162025181178,
|
| 1823 |
+
"learning_rate": 2.8193551976335525e-05,
|
| 1824 |
+
"loss": 0.4642,
|
| 1825 |
+
"num_tokens": 868277615.0,
|
| 1826 |
+
"step": 1135
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 1.5552523874488404,
|
| 1830 |
+
"grad_norm": 0.2344367845147417,
|
| 1831 |
+
"learning_rate": 2.8024428540288638e-05,
|
| 1832 |
+
"loss": 0.4735,
|
| 1833 |
+
"num_tokens": 872188105.0,
|
| 1834 |
+
"step": 1140
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 1.5620736698499318,
|
| 1838 |
+
"grad_norm": 0.2550702162457621,
|
| 1839 |
+
"learning_rate": 2.7855275452709094e-05,
|
| 1840 |
+
"loss": 0.4631,
|
| 1841 |
+
"num_tokens": 876046118.0,
|
| 1842 |
+
"step": 1145
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 1.5688949522510232,
|
| 1846 |
+
"grad_norm": 0.24686318609530264,
|
| 1847 |
+
"learning_rate": 2.76861022776231e-05,
|
| 1848 |
+
"loss": 0.462,
|
| 1849 |
+
"num_tokens": 879952254.0,
|
| 1850 |
+
"step": 1150
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 1.5757162346521145,
|
| 1854 |
+
"grad_norm": 0.2662544591075248,
|
| 1855 |
+
"learning_rate": 2.751691858019261e-05,
|
| 1856 |
+
"loss": 0.4659,
|
| 1857 |
+
"num_tokens": 883721474.0,
|
| 1858 |
+
"step": 1155
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 1.5825375170532059,
|
| 1862 |
+
"grad_norm": 0.254497506944391,
|
| 1863 |
+
"learning_rate": 2.7347733926174535e-05,
|
| 1864 |
+
"loss": 0.4823,
|
| 1865 |
+
"num_tokens": 887512011.0,
|
| 1866 |
+
"step": 1160
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 1.5893587994542973,
|
| 1870 |
+
"grad_norm": 0.24061168758830734,
|
| 1871 |
+
"learning_rate": 2.7178557881379845e-05,
|
| 1872 |
+
"loss": 0.4534,
|
| 1873 |
+
"num_tokens": 891332828.0,
|
| 1874 |
+
"step": 1165
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 1.5961800818553888,
|
| 1878 |
+
"grad_norm": 0.255092810360894,
|
| 1879 |
+
"learning_rate": 2.7009400011132784e-05,
|
| 1880 |
+
"loss": 0.4576,
|
| 1881 |
+
"num_tokens": 895212128.0,
|
| 1882 |
+
"step": 1170
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 1.6030013642564802,
|
| 1886 |
+
"grad_norm": 0.24118127655662266,
|
| 1887 |
+
"learning_rate": 2.6840269879729964e-05,
|
| 1888 |
+
"loss": 0.4585,
|
| 1889 |
+
"num_tokens": 899092598.0,
|
| 1890 |
+
"step": 1175
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 1.6098226466575716,
|
| 1894 |
+
"grad_norm": 0.2485285467248689,
|
| 1895 |
+
"learning_rate": 2.667117704989962e-05,
|
| 1896 |
+
"loss": 0.4625,
|
| 1897 |
+
"num_tokens": 902821411.0,
|
| 1898 |
+
"step": 1180
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 1.6166439290586632,
|
| 1902 |
+
"grad_norm": 0.24640583203612748,
|
| 1903 |
+
"learning_rate": 2.650213108226094e-05,
|
| 1904 |
+
"loss": 0.4605,
|
| 1905 |
+
"num_tokens": 906650186.0,
|
| 1906 |
+
"step": 1185
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 1.6234652114597545,
|
| 1910 |
+
"grad_norm": 0.23275634383221383,
|
| 1911 |
+
"learning_rate": 2.6333141534783507e-05,
|
| 1912 |
+
"loss": 0.464,
|
| 1913 |
+
"num_tokens": 910561700.0,
|
| 1914 |
+
"step": 1190
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.630286493860846,
|
| 1918 |
+
"grad_norm": 0.22346130896170255,
|
| 1919 |
+
"learning_rate": 2.6164217962246846e-05,
|
| 1920 |
+
"loss": 0.4574,
|
| 1921 |
+
"num_tokens": 914471979.0,
|
| 1922 |
+
"step": 1195
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 1.6371077762619373,
|
| 1926 |
+
"grad_norm": 0.23177866181041107,
|
| 1927 |
+
"learning_rate": 2.5995369915700242e-05,
|
| 1928 |
+
"loss": 0.4601,
|
| 1929 |
+
"num_tokens": 918407451.0,
|
| 1930 |
+
"step": 1200
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 1.6439290586630286,
|
| 1934 |
+
"grad_norm": 0.230861591331699,
|
| 1935 |
+
"learning_rate": 2.5826606941922664e-05,
|
| 1936 |
+
"loss": 0.4499,
|
| 1937 |
+
"num_tokens": 922149664.0,
|
| 1938 |
+
"step": 1205
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 1.65075034106412,
|
| 1942 |
+
"grad_norm": 0.2798858214328977,
|
| 1943 |
+
"learning_rate": 2.5657938582883034e-05,
|
| 1944 |
+
"loss": 0.4623,
|
| 1945 |
+
"num_tokens": 925906310.0,
|
| 1946 |
+
"step": 1210
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 1.6575716234652114,
|
| 1950 |
+
"grad_norm": 0.26225451170335046,
|
| 1951 |
+
"learning_rate": 2.548937437520067e-05,
|
| 1952 |
+
"loss": 0.4552,
|
| 1953 |
+
"num_tokens": 929886517.0,
|
| 1954 |
+
"step": 1215
|
| 1955 |
+
},
|
| 1956 |
+
{
|
| 1957 |
+
"epoch": 1.6643929058663027,
|
| 1958 |
+
"grad_norm": 0.2153042049132024,
|
| 1959 |
+
"learning_rate": 2.532092384960614e-05,
|
| 1960 |
+
"loss": 0.4477,
|
| 1961 |
+
"num_tokens": 933572845.0,
|
| 1962 |
+
"step": 1220
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 1.6712141882673943,
|
| 1966 |
+
"grad_norm": 0.2418303845011376,
|
| 1967 |
+
"learning_rate": 2.5152596530402327e-05,
|
| 1968 |
+
"loss": 0.4517,
|
| 1969 |
+
"num_tokens": 937133457.0,
|
| 1970 |
+
"step": 1225
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 1.6780354706684857,
|
| 1974 |
+
"grad_norm": 0.24947742050579547,
|
| 1975 |
+
"learning_rate": 2.4984401934925945e-05,
|
| 1976 |
+
"loss": 0.4548,
|
| 1977 |
+
"num_tokens": 941038842.0,
|
| 1978 |
+
"step": 1230
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 1.684856753069577,
|
| 1982 |
+
"grad_norm": 0.23458060058465702,
|
| 1983 |
+
"learning_rate": 2.4816349573009422e-05,
|
| 1984 |
+
"loss": 0.4626,
|
| 1985 |
+
"num_tokens": 944824346.0,
|
| 1986 |
+
"step": 1235
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 1.6916780354706686,
|
| 1990 |
+
"grad_norm": 0.24159560615393907,
|
| 1991 |
+
"learning_rate": 2.464844894644322e-05,
|
| 1992 |
+
"loss": 0.4525,
|
| 1993 |
+
"num_tokens": 948787821.0,
|
| 1994 |
+
"step": 1240
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 1.69849931787176,
|
| 1998 |
+
"grad_norm": 0.22102950015120187,
|
| 1999 |
+
"learning_rate": 2.448070954843857e-05,
|
| 2000 |
+
"loss": 0.4615,
|
| 2001 |
+
"num_tokens": 952546196.0,
|
| 2002 |
+
"step": 1245
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 1.7053206002728514,
|
| 2006 |
+
"grad_norm": 0.2321267387454469,
|
| 2007 |
+
"learning_rate": 2.431314086309075e-05,
|
| 2008 |
+
"loss": 0.4593,
|
| 2009 |
+
"num_tokens": 956413312.0,
|
| 2010 |
+
"step": 1250
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 1.7121418826739427,
|
| 2014 |
+
"grad_norm": 0.23287977913756222,
|
| 2015 |
+
"learning_rate": 2.4145752364842807e-05,
|
| 2016 |
+
"loss": 0.4537,
|
| 2017 |
+
"num_tokens": 960097484.0,
|
| 2018 |
+
"step": 1255
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 1.718963165075034,
|
| 2022 |
+
"grad_norm": 0.23742326211987602,
|
| 2023 |
+
"learning_rate": 2.397855351794992e-05,
|
| 2024 |
+
"loss": 0.4502,
|
| 2025 |
+
"num_tokens": 963903477.0,
|
| 2026 |
+
"step": 1260
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 1.7257844474761255,
|
| 2030 |
+
"grad_norm": 0.22718319203576953,
|
| 2031 |
+
"learning_rate": 2.381155377594424e-05,
|
| 2032 |
+
"loss": 0.4555,
|
| 2033 |
+
"num_tokens": 967716958.0,
|
| 2034 |
+
"step": 1265
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 1.7326057298772168,
|
| 2038 |
+
"grad_norm": 0.2415598883613876,
|
| 2039 |
+
"learning_rate": 2.3644762581100398e-05,
|
| 2040 |
+
"loss": 0.4396,
|
| 2041 |
+
"num_tokens": 971436672.0,
|
| 2042 |
+
"step": 1270
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 1.7394270122783082,
|
| 2046 |
+
"grad_norm": 0.2392682686531376,
|
| 2047 |
+
"learning_rate": 2.3478189363901653e-05,
|
| 2048 |
+
"loss": 0.4662,
|
| 2049 |
+
"num_tokens": 975282872.0,
|
| 2050 |
+
"step": 1275
|
| 2051 |
+
},
|
| 2052 |
+
{
|
| 2053 |
+
"epoch": 1.7462482946793996,
|
| 2054 |
+
"grad_norm": 0.225547635686384,
|
| 2055 |
+
"learning_rate": 2.3311843542506632e-05,
|
| 2056 |
+
"loss": 0.459,
|
| 2057 |
+
"num_tokens": 979113101.0,
|
| 2058 |
+
"step": 1280
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 1.7530695770804912,
|
| 2062 |
+
"grad_norm": 0.21735597497530967,
|
| 2063 |
+
"learning_rate": 2.3145734522216883e-05,
|
| 2064 |
+
"loss": 0.4538,
|
| 2065 |
+
"num_tokens": 983117582.0,
|
| 2066 |
+
"step": 1285
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 1.7598908594815825,
|
| 2070 |
+
"grad_norm": 0.24743369312514374,
|
| 2071 |
+
"learning_rate": 2.2979871694945038e-05,
|
| 2072 |
+
"loss": 0.4646,
|
| 2073 |
+
"num_tokens": 986965249.0,
|
| 2074 |
+
"step": 1290
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.766712141882674,
|
| 2078 |
+
"grad_norm": 0.22546136414855156,
|
| 2079 |
+
"learning_rate": 2.281426443868384e-05,
|
| 2080 |
+
"loss": 0.4574,
|
| 2081 |
+
"num_tokens": 990643231.0,
|
| 2082 |
+
"step": 1295
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.7735334242837655,
|
| 2086 |
+
"grad_norm": 0.22415419229501857,
|
| 2087 |
+
"learning_rate": 2.264892211697588e-05,
|
| 2088 |
+
"loss": 0.4492,
|
| 2089 |
+
"num_tokens": 994547720.0,
|
| 2090 |
+
"step": 1300
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 1.7803547066848568,
|
| 2094 |
+
"grad_norm": 0.22773316789517045,
|
| 2095 |
+
"learning_rate": 2.248385407838414e-05,
|
| 2096 |
+
"loss": 0.4513,
|
| 2097 |
+
"num_tokens": 998141405.0,
|
| 2098 |
+
"step": 1305
|
| 2099 |
+
},
|
| 2100 |
+
{
|
| 2101 |
+
"epoch": 1.7871759890859482,
|
| 2102 |
+
"grad_norm": 0.22216852340940998,
|
| 2103 |
+
"learning_rate": 2.2319069655963493e-05,
|
| 2104 |
+
"loss": 0.4488,
|
| 2105 |
+
"num_tokens": 1001828735.0,
|
| 2106 |
+
"step": 1310
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 1.7939972714870396,
|
| 2110 |
+
"grad_norm": 0.2266447362168477,
|
| 2111 |
+
"learning_rate": 2.2154578166732953e-05,
|
| 2112 |
+
"loss": 0.4621,
|
| 2113 |
+
"num_tokens": 1005562335.0,
|
| 2114 |
+
"step": 1315
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 1.800818553888131,
|
| 2118 |
+
"grad_norm": 0.23208103808871097,
|
| 2119 |
+
"learning_rate": 2.1990388911148912e-05,
|
| 2120 |
+
"loss": 0.455,
|
| 2121 |
+
"num_tokens": 1009141104.0,
|
| 2122 |
+
"step": 1320
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 1.8076398362892223,
|
| 2126 |
+
"grad_norm": 0.27534377132464494,
|
| 2127 |
+
"learning_rate": 2.182651117257925e-05,
|
| 2128 |
+
"loss": 0.4602,
|
| 2129 |
+
"num_tokens": 1012723565.0,
|
| 2130 |
+
"step": 1325
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.8144611186903137,
|
| 2134 |
+
"grad_norm": 0.2426506801004899,
|
| 2135 |
+
"learning_rate": 2.1662954216778496e-05,
|
| 2136 |
+
"loss": 0.4516,
|
| 2137 |
+
"num_tokens": 1016477651.0,
|
| 2138 |
+
"step": 1330
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 1.821282401091405,
|
| 2142 |
+
"grad_norm": 0.2473170968483256,
|
| 2143 |
+
"learning_rate": 2.14997272913639e-05,
|
| 2144 |
+
"loss": 0.4543,
|
| 2145 |
+
"num_tokens": 1020301965.0,
|
| 2146 |
+
"step": 1335
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 1.8281036834924966,
|
| 2150 |
+
"grad_norm": 0.22163949530755686,
|
| 2151 |
+
"learning_rate": 2.1336839625292582e-05,
|
| 2152 |
+
"loss": 0.4447,
|
| 2153 |
+
"num_tokens": 1024157287.0,
|
| 2154 |
+
"step": 1340
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 1.834924965893588,
|
| 2158 |
+
"grad_norm": 0.24698105825618707,
|
| 2159 |
+
"learning_rate": 2.1174300428339715e-05,
|
| 2160 |
+
"loss": 0.464,
|
| 2161 |
+
"num_tokens": 1027955809.0,
|
| 2162 |
+
"step": 1345
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 1.8417462482946794,
|
| 2166 |
+
"grad_norm": 0.21848253671188075,
|
| 2167 |
+
"learning_rate": 2.1012118890577787e-05,
|
| 2168 |
+
"loss": 0.4558,
|
| 2169 |
+
"num_tokens": 1031820433.0,
|
| 2170 |
+
"step": 1350
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 1.848567530695771,
|
| 2174 |
+
"grad_norm": 0.2263316143914313,
|
| 2175 |
+
"learning_rate": 2.085030418185703e-05,
|
| 2176 |
+
"loss": 0.4569,
|
| 2177 |
+
"num_tokens": 1035628174.0,
|
| 2178 |
+
"step": 1355
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 1.8553888130968623,
|
| 2182 |
+
"grad_norm": 0.23089585910089416,
|
| 2183 |
+
"learning_rate": 2.0688865451286888e-05,
|
| 2184 |
+
"loss": 0.4656,
|
| 2185 |
+
"num_tokens": 1039538259.0,
|
| 2186 |
+
"step": 1360
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 1.8622100954979537,
|
| 2190 |
+
"grad_norm": 0.23862099503139816,
|
| 2191 |
+
"learning_rate": 2.052781182671879e-05,
|
| 2192 |
+
"loss": 0.4541,
|
| 2193 |
+
"num_tokens": 1043323479.0,
|
| 2194 |
+
"step": 1365
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 1.869031377899045,
|
| 2198 |
+
"grad_norm": 0.22978276166805608,
|
| 2199 |
+
"learning_rate": 2.0367152414229983e-05,
|
| 2200 |
+
"loss": 0.4599,
|
| 2201 |
+
"num_tokens": 1047143725.0,
|
| 2202 |
+
"step": 1370
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 1.8758526603001364,
|
| 2206 |
+
"grad_norm": 0.22012840335234812,
|
| 2207 |
+
"learning_rate": 2.0206896297608747e-05,
|
| 2208 |
+
"loss": 0.4548,
|
| 2209 |
+
"num_tokens": 1051058729.0,
|
| 2210 |
+
"step": 1375
|
| 2211 |
+
},
|
| 2212 |
+
{
|
| 2213 |
+
"epoch": 1.8826739427012278,
|
| 2214 |
+
"grad_norm": 0.23829493995999024,
|
| 2215 |
+
"learning_rate": 2.004705253784071e-05,
|
| 2216 |
+
"loss": 0.4498,
|
| 2217 |
+
"num_tokens": 1054707187.0,
|
| 2218 |
+
"step": 1380
|
| 2219 |
+
},
|
| 2220 |
+
{
|
| 2221 |
+
"epoch": 1.8894952251023192,
|
| 2222 |
+
"grad_norm": 0.2211953633014594,
|
| 2223 |
+
"learning_rate": 1.9887630172596605e-05,
|
| 2224 |
+
"loss": 0.4494,
|
| 2225 |
+
"num_tokens": 1058505830.0,
|
| 2226 |
+
"step": 1385
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 1.8963165075034105,
|
| 2230 |
+
"grad_norm": 0.2154212278200008,
|
| 2231 |
+
"learning_rate": 1.972863821572122e-05,
|
| 2232 |
+
"loss": 0.4434,
|
| 2233 |
+
"num_tokens": 1062382498.0,
|
| 2234 |
+
"step": 1390
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 1.9031377899045019,
|
| 2238 |
+
"grad_norm": 0.22559761404121423,
|
| 2239 |
+
"learning_rate": 1.957008565672378e-05,
|
| 2240 |
+
"loss": 0.4612,
|
| 2241 |
+
"num_tokens": 1066230098.0,
|
| 2242 |
+
"step": 1395
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 1.9099590723055935,
|
| 2246 |
+
"grad_norm": 0.2241792422987182,
|
| 2247 |
+
"learning_rate": 1.9411981460269668e-05,
|
| 2248 |
+
"loss": 0.4519,
|
| 2249 |
+
"num_tokens": 1070100345.0,
|
| 2250 |
+
"step": 1400
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 1.9167803547066848,
|
| 2254 |
+
"grad_norm": 0.22521629892381584,
|
| 2255 |
+
"learning_rate": 1.9254334565673565e-05,
|
| 2256 |
+
"loss": 0.4513,
|
| 2257 |
+
"num_tokens": 1073919676.0,
|
| 2258 |
+
"step": 1405
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 1.9236016371077762,
|
| 2262 |
+
"grad_norm": 0.22286740780536643,
|
| 2263 |
+
"learning_rate": 1.9097153886393988e-05,
|
| 2264 |
+
"loss": 0.4578,
|
| 2265 |
+
"num_tokens": 1077765957.0,
|
| 2266 |
+
"step": 1410
|
| 2267 |
+
},
|
| 2268 |
+
{
|
| 2269 |
+
"epoch": 1.9304229195088678,
|
| 2270 |
+
"grad_norm": 0.2320954952638967,
|
| 2271 |
+
"learning_rate": 1.894044830952934e-05,
|
| 2272 |
+
"loss": 0.4478,
|
| 2273 |
+
"num_tokens": 1081540834.0,
|
| 2274 |
+
"step": 1415
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 1.9372442019099592,
|
| 2278 |
+
"grad_norm": 0.22171734630993664,
|
| 2279 |
+
"learning_rate": 1.8784226695315438e-05,
|
| 2280 |
+
"loss": 0.4627,
|
| 2281 |
+
"num_tokens": 1085254168.0,
|
| 2282 |
+
"step": 1420
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 1.9440654843110505,
|
| 2286 |
+
"grad_norm": 0.22484798749427884,
|
| 2287 |
+
"learning_rate": 1.862849787662452e-05,
|
| 2288 |
+
"loss": 0.4556,
|
| 2289 |
+
"num_tokens": 1089204273.0,
|
| 2290 |
+
"step": 1425
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 1.950886766712142,
|
| 2294 |
+
"grad_norm": 0.2288368074532154,
|
| 2295 |
+
"learning_rate": 1.8473270658465818e-05,
|
| 2296 |
+
"loss": 0.4544,
|
| 2297 |
+
"num_tokens": 1093059842.0,
|
| 2298 |
+
"step": 1430
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 1.9577080491132333,
|
| 2302 |
+
"grad_norm": 0.21112016863141217,
|
| 2303 |
+
"learning_rate": 1.8318553817487817e-05,
|
| 2304 |
+
"loss": 0.4598,
|
| 2305 |
+
"num_tokens": 1096876504.0,
|
| 2306 |
+
"step": 1435
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 1.9645293315143246,
|
| 2310 |
+
"grad_norm": 0.21943606003762398,
|
| 2311 |
+
"learning_rate": 1.8164356101481867e-05,
|
| 2312 |
+
"loss": 0.448,
|
| 2313 |
+
"num_tokens": 1100833806.0,
|
| 2314 |
+
"step": 1440
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 1.971350613915416,
|
| 2318 |
+
"grad_norm": 0.22709413239779286,
|
| 2319 |
+
"learning_rate": 1.801068622888768e-05,
|
| 2320 |
+
"loss": 0.4488,
|
| 2321 |
+
"num_tokens": 1104630572.0,
|
| 2322 |
+
"step": 1445
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 1.9781718963165074,
|
| 2326 |
+
"grad_norm": 0.21353005961467678,
|
| 2327 |
+
"learning_rate": 1.7857552888300354e-05,
|
| 2328 |
+
"loss": 0.4437,
|
| 2329 |
+
"num_tokens": 1108388900.0,
|
| 2330 |
+
"step": 1450
|
| 2331 |
+
},
|
| 2332 |
+
{
|
| 2333 |
+
"epoch": 1.984993178717599,
|
| 2334 |
+
"grad_norm": 0.22771273689058968,
|
| 2335 |
+
"learning_rate": 1.770496473797914e-05,
|
| 2336 |
+
"loss": 0.4588,
|
| 2337 |
+
"num_tokens": 1112326009.0,
|
| 2338 |
+
"step": 1455
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 1.9918144611186903,
|
| 2342 |
+
"grad_norm": 0.21294407368974086,
|
| 2343 |
+
"learning_rate": 1.755293040535783e-05,
|
| 2344 |
+
"loss": 0.4556,
|
| 2345 |
+
"num_tokens": 1116108995.0,
|
| 2346 |
+
"step": 1460
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 1.9986357435197817,
|
| 2350 |
+
"grad_norm": 0.22052518377500083,
|
| 2351 |
+
"learning_rate": 1.7401458486557027e-05,
|
| 2352 |
+
"loss": 0.4455,
|
| 2353 |
+
"num_tokens": 1119855365.0,
|
| 2354 |
+
"step": 1465
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 2.0054570259208733,
|
| 2358 |
+
"grad_norm": 0.26970555548774344,
|
| 2359 |
+
"learning_rate": 1.7250557545898104e-05,
|
| 2360 |
+
"loss": 0.423,
|
| 2361 |
+
"num_tokens": 1123684625.0,
|
| 2362 |
+
"step": 1470
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 2.0122783083219646,
|
| 2366 |
+
"grad_norm": 0.24152497077350163,
|
| 2367 |
+
"learning_rate": 1.710023611541892e-05,
|
| 2368 |
+
"loss": 0.4172,
|
| 2369 |
+
"num_tokens": 1127572305.0,
|
| 2370 |
+
"step": 1475
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.019099590723056,
|
| 2374 |
+
"grad_norm": 0.23841549865625594,
|
| 2375 |
+
"learning_rate": 1.695050269439148e-05,
|
| 2376 |
+
"loss": 0.4266,
|
| 2377 |
+
"num_tokens": 1131358945.0,
|
| 2378 |
+
"step": 1480
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 2.0259208731241474,
|
| 2382 |
+
"grad_norm": 0.22139442741982837,
|
| 2383 |
+
"learning_rate": 1.6801365748841326e-05,
|
| 2384 |
+
"loss": 0.4213,
|
| 2385 |
+
"num_tokens": 1135163182.0,
|
| 2386 |
+
"step": 1485
|
| 2387 |
+
},
|
| 2388 |
+
{
|
| 2389 |
+
"epoch": 2.0327421555252387,
|
| 2390 |
+
"grad_norm": 0.2034705392819918,
|
| 2391 |
+
"learning_rate": 1.665283371106891e-05,
|
| 2392 |
+
"loss": 0.4335,
|
| 2393 |
+
"num_tokens": 1139156691.0,
|
| 2394 |
+
"step": 1490
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 2.03956343792633,
|
| 2398 |
+
"grad_norm": 0.21895722839318674,
|
| 2399 |
+
"learning_rate": 1.6504914979172787e-05,
|
| 2400 |
+
"loss": 0.4229,
|
| 2401 |
+
"num_tokens": 1142963803.0,
|
| 2402 |
+
"step": 1495
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 2.0463847203274215,
|
| 2406 |
+
"grad_norm": 0.2291442238701152,
|
| 2407 |
+
"learning_rate": 1.6357617916574814e-05,
|
| 2408 |
+
"loss": 0.41,
|
| 2409 |
+
"num_tokens": 1146661529.0,
|
| 2410 |
+
"step": 1500
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 2.053206002728513,
|
| 2414 |
+
"grad_norm": 0.23049833465167152,
|
| 2415 |
+
"learning_rate": 1.6210950851547217e-05,
|
| 2416 |
+
"loss": 0.4301,
|
| 2417 |
+
"num_tokens": 1150631571.0,
|
| 2418 |
+
"step": 1505
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 2.060027285129604,
|
| 2422 |
+
"grad_norm": 0.22806866165986595,
|
| 2423 |
+
"learning_rate": 1.6064922076741808e-05,
|
| 2424 |
+
"loss": 0.4189,
|
| 2425 |
+
"num_tokens": 1154372520.0,
|
| 2426 |
+
"step": 1510
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.0668485675306956,
|
| 2430 |
+
"grad_norm": 0.22193454737326296,
|
| 2431 |
+
"learning_rate": 1.5919539848720995e-05,
|
| 2432 |
+
"loss": 0.4245,
|
| 2433 |
+
"num_tokens": 1158377799.0,
|
| 2434 |
+
"step": 1515
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 2.0736698499317874,
|
| 2438 |
+
"grad_norm": 0.22994821375176555,
|
| 2439 |
+
"learning_rate": 1.5774812387491054e-05,
|
| 2440 |
+
"loss": 0.4274,
|
| 2441 |
+
"num_tokens": 1162137345.0,
|
| 2442 |
+
"step": 1520
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 2.0804911323328787,
|
| 2446 |
+
"grad_norm": 0.26138494069692764,
|
| 2447 |
+
"learning_rate": 1.5630747876037284e-05,
|
| 2448 |
+
"loss": 0.4268,
|
| 2449 |
+
"num_tokens": 1165931384.0,
|
| 2450 |
+
"step": 1525
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 2.08731241473397,
|
| 2454 |
+
"grad_norm": 0.21924047377156536,
|
| 2455 |
+
"learning_rate": 1.548735445986139e-05,
|
| 2456 |
+
"loss": 0.4192,
|
| 2457 |
+
"num_tokens": 1169758424.0,
|
| 2458 |
+
"step": 1530
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 2.0941336971350615,
|
| 2462 |
+
"grad_norm": 0.2521336569888774,
|
| 2463 |
+
"learning_rate": 1.5344640246520908e-05,
|
| 2464 |
+
"loss": 0.4247,
|
| 2465 |
+
"num_tokens": 1173501069.0,
|
| 2466 |
+
"step": 1535
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 2.100954979536153,
|
| 2470 |
+
"grad_norm": 0.23766806626207862,
|
| 2471 |
+
"learning_rate": 1.5202613305170793e-05,
|
| 2472 |
+
"loss": 0.4315,
|
| 2473 |
+
"num_tokens": 1177335233.0,
|
| 2474 |
+
"step": 1540
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 2.107776261937244,
|
| 2478 |
+
"grad_norm": 0.21367914174173042,
|
| 2479 |
+
"learning_rate": 1.5061281666107208e-05,
|
| 2480 |
+
"loss": 0.4204,
|
| 2481 |
+
"num_tokens": 1181125439.0,
|
| 2482 |
+
"step": 1545
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.1145975443383356,
|
| 2486 |
+
"grad_norm": 0.2694569747992092,
|
| 2487 |
+
"learning_rate": 1.4920653320313457e-05,
|
| 2488 |
+
"loss": 0.4303,
|
| 2489 |
+
"num_tokens": 1185004310.0,
|
| 2490 |
+
"step": 1550
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 2.121418826739427,
|
| 2494 |
+
"grad_norm": 0.2429660678216922,
|
| 2495 |
+
"learning_rate": 1.4780736219008174e-05,
|
| 2496 |
+
"loss": 0.4352,
|
| 2497 |
+
"num_tokens": 1188730220.0,
|
| 2498 |
+
"step": 1555
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 2.1282401091405183,
|
| 2502 |
+
"grad_norm": 0.21473274786268579,
|
| 2503 |
+
"learning_rate": 1.464153827319579e-05,
|
| 2504 |
+
"loss": 0.4145,
|
| 2505 |
+
"num_tokens": 1192452885.0,
|
| 2506 |
+
"step": 1560
|
| 2507 |
+
},
|
| 2508 |
+
{
|
| 2509 |
+
"epoch": 2.1350613915416097,
|
| 2510 |
+
"grad_norm": 0.22760086194375342,
|
| 2511 |
+
"learning_rate": 1.4503067353219185e-05,
|
| 2512 |
+
"loss": 0.421,
|
| 2513 |
+
"num_tokens": 1196235271.0,
|
| 2514 |
+
"step": 1565
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 2.141882673942701,
|
| 2518 |
+
"grad_norm": 0.23938548757452846,
|
| 2519 |
+
"learning_rate": 1.436533128831476e-05,
|
| 2520 |
+
"loss": 0.4177,
|
| 2521 |
+
"num_tokens": 1200064083.0,
|
| 2522 |
+
"step": 1570
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 2.148703956343793,
|
| 2526 |
+
"grad_norm": 0.2116741668932284,
|
| 2527 |
+
"learning_rate": 1.4228337866169687e-05,
|
| 2528 |
+
"loss": 0.4253,
|
| 2529 |
+
"num_tokens": 1203975750.0,
|
| 2530 |
+
"step": 1575
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 2.155525238744884,
|
| 2534 |
+
"grad_norm": 0.22023980635986193,
|
| 2535 |
+
"learning_rate": 1.4092094832481664e-05,
|
| 2536 |
+
"loss": 0.429,
|
| 2537 |
+
"num_tokens": 1207941601.0,
|
| 2538 |
+
"step": 1580
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.1623465211459756,
|
| 2542 |
+
"grad_norm": 0.22509953830623908,
|
| 2543 |
+
"learning_rate": 1.3956609890520919e-05,
|
| 2544 |
+
"loss": 0.419,
|
| 2545 |
+
"num_tokens": 1211722253.0,
|
| 2546 |
+
"step": 1585
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 2.169167803547067,
|
| 2550 |
+
"grad_norm": 0.22352404953702504,
|
| 2551 |
+
"learning_rate": 1.3821890700694668e-05,
|
| 2552 |
+
"loss": 0.4258,
|
| 2553 |
+
"num_tokens": 1215472991.0,
|
| 2554 |
+
"step": 1590
|
| 2555 |
+
},
|
| 2556 |
+
{
|
| 2557 |
+
"epoch": 2.1759890859481583,
|
| 2558 |
+
"grad_norm": 0.2181159636375956,
|
| 2559 |
+
"learning_rate": 1.3687944880114034e-05,
|
| 2560 |
+
"loss": 0.4297,
|
| 2561 |
+
"num_tokens": 1219314367.0,
|
| 2562 |
+
"step": 1595
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 2.1828103683492497,
|
| 2566 |
+
"grad_norm": 0.2306072063541757,
|
| 2567 |
+
"learning_rate": 1.3554780002163312e-05,
|
| 2568 |
+
"loss": 0.4312,
|
| 2569 |
+
"num_tokens": 1223192538.0,
|
| 2570 |
+
"step": 1600
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 2.189631650750341,
|
| 2574 |
+
"grad_norm": 0.2283007086354253,
|
| 2575 |
+
"learning_rate": 1.3422403596071778e-05,
|
| 2576 |
+
"loss": 0.4081,
|
| 2577 |
+
"num_tokens": 1226911643.0,
|
| 2578 |
+
"step": 1605
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 2.1964529331514324,
|
| 2582 |
+
"grad_norm": 0.21220002250579495,
|
| 2583 |
+
"learning_rate": 1.3290823146488029e-05,
|
| 2584 |
+
"loss": 0.4233,
|
| 2585 |
+
"num_tokens": 1230700907.0,
|
| 2586 |
+
"step": 1610
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 2.203274215552524,
|
| 2590 |
+
"grad_norm": 0.22244492420199785,
|
| 2591 |
+
"learning_rate": 1.3160046093056732e-05,
|
| 2592 |
+
"loss": 0.4271,
|
| 2593 |
+
"num_tokens": 1234481818.0,
|
| 2594 |
+
"step": 1615
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.210095497953615,
|
| 2598 |
+
"grad_norm": 0.21042010937277456,
|
| 2599 |
+
"learning_rate": 1.3030079829998027e-05,
|
| 2600 |
+
"loss": 0.426,
|
| 2601 |
+
"num_tokens": 1238353778.0,
|
| 2602 |
+
"step": 1620
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 2.2169167803547065,
|
| 2606 |
+
"grad_norm": 0.21189963519565957,
|
| 2607 |
+
"learning_rate": 1.2900931705689424e-05,
|
| 2608 |
+
"loss": 0.4235,
|
| 2609 |
+
"num_tokens": 1242159504.0,
|
| 2610 |
+
"step": 1625
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"epoch": 2.223738062755798,
|
| 2614 |
+
"grad_norm": 0.21604962976902467,
|
| 2615 |
+
"learning_rate": 1.2772609022250326e-05,
|
| 2616 |
+
"loss": 0.4249,
|
| 2617 |
+
"num_tokens": 1246049292.0,
|
| 2618 |
+
"step": 1630
|
| 2619 |
+
},
|
| 2620 |
+
{
|
| 2621 |
+
"epoch": 2.2305593451568897,
|
| 2622 |
+
"grad_norm": 0.21749329314406493,
|
| 2623 |
+
"learning_rate": 1.26451190351292e-05,
|
| 2624 |
+
"loss": 0.4338,
|
| 2625 |
+
"num_tokens": 1249978353.0,
|
| 2626 |
+
"step": 1635
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 2.237380627557981,
|
| 2630 |
+
"grad_norm": 0.22645274952882502,
|
| 2631 |
+
"learning_rate": 1.2518468952693285e-05,
|
| 2632 |
+
"loss": 0.4218,
|
| 2633 |
+
"num_tokens": 1253553694.0,
|
| 2634 |
+
"step": 1640
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 2.2442019099590724,
|
| 2638 |
+
"grad_norm": 0.22556232039774654,
|
| 2639 |
+
"learning_rate": 1.2392665935821079e-05,
|
| 2640 |
+
"loss": 0.4253,
|
| 2641 |
+
"num_tokens": 1257413581.0,
|
| 2642 |
+
"step": 1645
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 2.251023192360164,
|
| 2646 |
+
"grad_norm": 0.22784202665727274,
|
| 2647 |
+
"learning_rate": 1.2267717097497444e-05,
|
| 2648 |
+
"loss": 0.4286,
|
| 2649 |
+
"num_tokens": 1261372211.0,
|
| 2650 |
+
"step": 1650
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.257844474761255,
|
| 2654 |
+
"grad_norm": 0.21889114756092548,
|
| 2655 |
+
"learning_rate": 1.2143629502411402e-05,
|
| 2656 |
+
"loss": 0.4237,
|
| 2657 |
+
"num_tokens": 1265194682.0,
|
| 2658 |
+
"step": 1655
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 2.2646657571623465,
|
| 2662 |
+
"grad_norm": 0.2157600429162021,
|
| 2663 |
+
"learning_rate": 1.2020410166556752e-05,
|
| 2664 |
+
"loss": 0.4219,
|
| 2665 |
+
"num_tokens": 1268958243.0,
|
| 2666 |
+
"step": 1660
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 2.271487039563438,
|
| 2670 |
+
"grad_norm": 0.2086503043234301,
|
| 2671 |
+
"learning_rate": 1.1898066056835314e-05,
|
| 2672 |
+
"loss": 0.4103,
|
| 2673 |
+
"num_tokens": 1272721552.0,
|
| 2674 |
+
"step": 1665
|
| 2675 |
+
},
|
| 2676 |
+
{
|
| 2677 |
+
"epoch": 2.2783083219645293,
|
| 2678 |
+
"grad_norm": 0.22792328479358706,
|
| 2679 |
+
"learning_rate": 1.1776604090663087e-05,
|
| 2680 |
+
"loss": 0.4197,
|
| 2681 |
+
"num_tokens": 1276637460.0,
|
| 2682 |
+
"step": 1670
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 2.2851296043656206,
|
| 2686 |
+
"grad_norm": 0.23127868841840266,
|
| 2687 |
+
"learning_rate": 1.1656031135579065e-05,
|
| 2688 |
+
"loss": 0.4279,
|
| 2689 |
+
"num_tokens": 1280437716.0,
|
| 2690 |
+
"step": 1675
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 2.291950886766712,
|
| 2694 |
+
"grad_norm": 0.20172649642338233,
|
| 2695 |
+
"learning_rate": 1.1536354008856982e-05,
|
| 2696 |
+
"loss": 0.4227,
|
| 2697 |
+
"num_tokens": 1284303404.0,
|
| 2698 |
+
"step": 1680
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 2.2987721691678034,
|
| 2702 |
+
"grad_norm": 0.20508345669716468,
|
| 2703 |
+
"learning_rate": 1.1417579477119871e-05,
|
| 2704 |
+
"loss": 0.4207,
|
| 2705 |
+
"num_tokens": 1288266060.0,
|
| 2706 |
+
"step": 1685
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.305593451568895,
|
| 2710 |
+
"grad_norm": 0.21328294309871434,
|
| 2711 |
+
"learning_rate": 1.129971425595742e-05,
|
| 2712 |
+
"loss": 0.4265,
|
| 2713 |
+
"num_tokens": 1292088844.0,
|
| 2714 |
+
"step": 1690
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 2.3124147339699865,
|
| 2718 |
+
"grad_norm": 0.2247060755563071,
|
| 2719 |
+
"learning_rate": 1.1182765009546329e-05,
|
| 2720 |
+
"loss": 0.4188,
|
| 2721 |
+
"num_tokens": 1295907279.0,
|
| 2722 |
+
"step": 1695
|
| 2723 |
+
},
|
| 2724 |
+
{
|
| 2725 |
+
"epoch": 2.319236016371078,
|
| 2726 |
+
"grad_norm": 0.22177341155852484,
|
| 2727 |
+
"learning_rate": 1.1066738350273485e-05,
|
| 2728 |
+
"loss": 0.4134,
|
| 2729 |
+
"num_tokens": 1299636811.0,
|
| 2730 |
+
"step": 1700
|
| 2731 |
+
},
|
| 2732 |
+
{
|
| 2733 |
+
"epoch": 2.3260572987721693,
|
| 2734 |
+
"grad_norm": 0.2132614711559736,
|
| 2735 |
+
"learning_rate": 1.0951640838362069e-05,
|
| 2736 |
+
"loss": 0.4284,
|
| 2737 |
+
"num_tokens": 1303503660.0,
|
| 2738 |
+
"step": 1705
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 2.3328785811732606,
|
| 2742 |
+
"grad_norm": 0.21443037530341233,
|
| 2743 |
+
"learning_rate": 1.0837478981500704e-05,
|
| 2744 |
+
"loss": 0.4288,
|
| 2745 |
+
"num_tokens": 1307218513.0,
|
| 2746 |
+
"step": 1710
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.339699863574352,
|
| 2750 |
+
"grad_norm": 0.19933486371855674,
|
| 2751 |
+
"learning_rate": 1.072425923447542e-05,
|
| 2752 |
+
"loss": 0.4159,
|
| 2753 |
+
"num_tokens": 1311018055.0,
|
| 2754 |
+
"step": 1715
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 2.3465211459754434,
|
| 2758 |
+
"grad_norm": 0.21852140800024134,
|
| 2759 |
+
"learning_rate": 1.0611987998804782e-05,
|
| 2760 |
+
"loss": 0.418,
|
| 2761 |
+
"num_tokens": 1314633285.0,
|
| 2762 |
+
"step": 1720
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.3533424283765347,
|
| 2766 |
+
"grad_norm": 0.2102437544749993,
|
| 2767 |
+
"learning_rate": 1.0500671622377871e-05,
|
| 2768 |
+
"loss": 0.4172,
|
| 2769 |
+
"num_tokens": 1318426856.0,
|
| 2770 |
+
"step": 1725
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 2.360163710777626,
|
| 2774 |
+
"grad_norm": 0.2199849127756761,
|
| 2775 |
+
"learning_rate": 1.0390316399095412e-05,
|
| 2776 |
+
"loss": 0.4173,
|
| 2777 |
+
"num_tokens": 1322084417.0,
|
| 2778 |
+
"step": 1730
|
| 2779 |
+
},
|
| 2780 |
+
{
|
| 2781 |
+
"epoch": 2.3669849931787175,
|
| 2782 |
+
"grad_norm": 0.19943717767548116,
|
| 2783 |
+
"learning_rate": 1.0280928568513911e-05,
|
| 2784 |
+
"loss": 0.4217,
|
| 2785 |
+
"num_tokens": 1325960721.0,
|
| 2786 |
+
"step": 1735
|
| 2787 |
+
},
|
| 2788 |
+
{
|
| 2789 |
+
"epoch": 2.373806275579809,
|
| 2790 |
+
"grad_norm": 0.19119426504413253,
|
| 2791 |
+
"learning_rate": 1.0172514315492849e-05,
|
| 2792 |
+
"loss": 0.4099,
|
| 2793 |
+
"num_tokens": 1329717684.0,
|
| 2794 |
+
"step": 1740
|
| 2795 |
+
},
|
| 2796 |
+
{
|
| 2797 |
+
"epoch": 2.3806275579809,
|
| 2798 |
+
"grad_norm": 0.21003875387253998,
|
| 2799 |
+
"learning_rate": 1.0065079769844999e-05,
|
| 2800 |
+
"loss": 0.4142,
|
| 2801 |
+
"num_tokens": 1333545659.0,
|
| 2802 |
+
"step": 1745
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 2.3874488403819916,
|
| 2806 |
+
"grad_norm": 0.20811397589721858,
|
| 2807 |
+
"learning_rate": 9.95863100598983e-06,
|
| 2808 |
+
"loss": 0.42,
|
| 2809 |
+
"num_tokens": 1337260747.0,
|
| 2810 |
+
"step": 1750
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 2.3942701227830834,
|
| 2814 |
+
"grad_norm": 0.22091049123979542,
|
| 2815 |
+
"learning_rate": 9.853174042610075e-06,
|
| 2816 |
+
"loss": 0.4217,
|
| 2817 |
+
"num_tokens": 1341043858.0,
|
| 2818 |
+
"step": 1755
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.4010914051841747,
|
| 2822 |
+
"grad_norm": 0.21152589956444964,
|
| 2823 |
+
"learning_rate": 9.748714842311423e-06,
|
| 2824 |
+
"loss": 0.4326,
|
| 2825 |
+
"num_tokens": 1344938218.0,
|
| 2826 |
+
"step": 1760
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 2.407912687585266,
|
| 2830 |
+
"grad_norm": 0.21237031984909194,
|
| 2831 |
+
"learning_rate": 9.64525931128537e-06,
|
| 2832 |
+
"loss": 0.4303,
|
| 2833 |
+
"num_tokens": 1348788919.0,
|
| 2834 |
+
"step": 1765
|
| 2835 |
+
},
|
| 2836 |
+
{
|
| 2837 |
+
"epoch": 2.4147339699863575,
|
| 2838 |
+
"grad_norm": 0.21191153462686857,
|
| 2839 |
+
"learning_rate": 9.542813298975298e-06,
|
| 2840 |
+
"loss": 0.4217,
|
| 2841 |
+
"num_tokens": 1352691877.0,
|
| 2842 |
+
"step": 1770
|
| 2843 |
+
},
|
| 2844 |
+
{
|
| 2845 |
+
"epoch": 2.421555252387449,
|
| 2846 |
+
"grad_norm": 0.20947444260690812,
|
| 2847 |
+
"learning_rate": 9.441382597745757e-06,
|
| 2848 |
+
"loss": 0.4236,
|
| 2849 |
+
"num_tokens": 1356448019.0,
|
| 2850 |
+
"step": 1775
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 2.42837653478854,
|
| 2854 |
+
"grad_norm": 0.21756200845329884,
|
| 2855 |
+
"learning_rate": 9.34097294255492e-06,
|
| 2856 |
+
"loss": 0.4277,
|
| 2857 |
+
"num_tokens": 1360291674.0,
|
| 2858 |
+
"step": 1780
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 2.4351978171896316,
|
| 2862 |
+
"grad_norm": 0.2107069815071612,
|
| 2863 |
+
"learning_rate": 9.241590010630374e-06,
|
| 2864 |
+
"loss": 0.4187,
|
| 2865 |
+
"num_tokens": 1364175739.0,
|
| 2866 |
+
"step": 1785
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 2.442019099590723,
|
| 2870 |
+
"grad_norm": 0.2057554247087737,
|
| 2871 |
+
"learning_rate": 9.143239421148079e-06,
|
| 2872 |
+
"loss": 0.4269,
|
| 2873 |
+
"num_tokens": 1368149466.0,
|
| 2874 |
+
"step": 1790
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 2.4488403819918143,
|
| 2878 |
+
"grad_norm": 0.2011574588937702,
|
| 2879 |
+
"learning_rate": 9.045926734914695e-06,
|
| 2880 |
+
"loss": 0.4207,
|
| 2881 |
+
"num_tokens": 1371952154.0,
|
| 2882 |
+
"step": 1795
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 2.4556616643929057,
|
| 2886 |
+
"grad_norm": 0.216207905344711,
|
| 2887 |
+
"learning_rate": 8.949657454053145e-06,
|
| 2888 |
+
"loss": 0.4236,
|
| 2889 |
+
"num_tokens": 1375728846.0,
|
| 2890 |
+
"step": 1800
|
| 2891 |
+
},
|
| 2892 |
+
{
|
| 2893 |
+
"epoch": 2.4624829467939975,
|
| 2894 |
+
"grad_norm": 0.1996476815957156,
|
| 2895 |
+
"learning_rate": 8.85443702169152e-06,
|
| 2896 |
+
"loss": 0.4164,
|
| 2897 |
+
"num_tokens": 1379383893.0,
|
| 2898 |
+
"step": 1805
|
| 2899 |
+
},
|
| 2900 |
+
{
|
| 2901 |
+
"epoch": 2.469304229195089,
|
| 2902 |
+
"grad_norm": 0.20698697867899082,
|
| 2903 |
+
"learning_rate": 8.760270821655362e-06,
|
| 2904 |
+
"loss": 0.4235,
|
| 2905 |
+
"num_tokens": 1383235087.0,
|
| 2906 |
+
"step": 1810
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 2.47612551159618,
|
| 2910 |
+
"grad_norm": 0.21034544532672167,
|
| 2911 |
+
"learning_rate": 8.667164178163201e-06,
|
| 2912 |
+
"loss": 0.4137,
|
| 2913 |
+
"num_tokens": 1387022452.0,
|
| 2914 |
+
"step": 1815
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 2.4829467939972716,
|
| 2918 |
+
"grad_norm": 0.19932543087987506,
|
| 2919 |
+
"learning_rate": 8.575122355525544e-06,
|
| 2920 |
+
"loss": 0.4204,
|
| 2921 |
+
"num_tokens": 1390822247.0,
|
| 2922 |
+
"step": 1820
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 2.489768076398363,
|
| 2926 |
+
"grad_norm": 0.1932802172505174,
|
| 2927 |
+
"learning_rate": 8.484150557847257e-06,
|
| 2928 |
+
"loss": 0.4194,
|
| 2929 |
+
"num_tokens": 1394564010.0,
|
| 2930 |
+
"step": 1825
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 2.4965893587994543,
|
| 2934 |
+
"grad_norm": 0.203871620759725,
|
| 2935 |
+
"learning_rate": 8.394253928733261e-06,
|
| 2936 |
+
"loss": 0.4153,
|
| 2937 |
+
"num_tokens": 1398382007.0,
|
| 2938 |
+
"step": 1830
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 2.5034106412005457,
|
| 2942 |
+
"grad_norm": 0.22291797700816554,
|
| 2943 |
+
"learning_rate": 8.305437550997778e-06,
|
| 2944 |
+
"loss": 0.421,
|
| 2945 |
+
"num_tokens": 1402127581.0,
|
| 2946 |
+
"step": 1835
|
| 2947 |
+
},
|
| 2948 |
+
{
|
| 2949 |
+
"epoch": 2.510231923601637,
|
| 2950 |
+
"grad_norm": 0.20344587634630842,
|
| 2951 |
+
"learning_rate": 8.217706446376874e-06,
|
| 2952 |
+
"loss": 0.4275,
|
| 2953 |
+
"num_tokens": 1406014675.0,
|
| 2954 |
+
"step": 1840
|
| 2955 |
+
},
|
| 2956 |
+
{
|
| 2957 |
+
"epoch": 2.5170532060027284,
|
| 2958 |
+
"grad_norm": 0.20401372830464223,
|
| 2959 |
+
"learning_rate": 8.131065575244603e-06,
|
| 2960 |
+
"loss": 0.4183,
|
| 2961 |
+
"num_tokens": 1409722672.0,
|
| 2962 |
+
"step": 1845
|
| 2963 |
+
},
|
| 2964 |
+
{
|
| 2965 |
+
"epoch": 2.52387448840382,
|
| 2966 |
+
"grad_norm": 0.20877894754217016,
|
| 2967 |
+
"learning_rate": 8.045519836332475e-06,
|
| 2968 |
+
"loss": 0.4267,
|
| 2969 |
+
"num_tokens": 1413666158.0,
|
| 2970 |
+
"step": 1850
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 2.530695770804911,
|
| 2974 |
+
"grad_norm": 0.1998080825840693,
|
| 2975 |
+
"learning_rate": 7.961074066452506e-06,
|
| 2976 |
+
"loss": 0.4281,
|
| 2977 |
+
"num_tokens": 1417402571.0,
|
| 2978 |
+
"step": 1855
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 2.5375170532060025,
|
| 2982 |
+
"grad_norm": 0.19584868512472456,
|
| 2983 |
+
"learning_rate": 7.877733040223775e-06,
|
| 2984 |
+
"loss": 0.4171,
|
| 2985 |
+
"num_tokens": 1421340412.0,
|
| 2986 |
+
"step": 1860
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 2.544338335607094,
|
| 2990 |
+
"grad_norm": 0.19514705966388332,
|
| 2991 |
+
"learning_rate": 7.795501469802402e-06,
|
| 2992 |
+
"loss": 0.4256,
|
| 2993 |
+
"num_tokens": 1425199953.0,
|
| 2994 |
+
"step": 1865
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 2.5511596180081857,
|
| 2998 |
+
"grad_norm": 0.1995417719559235,
|
| 2999 |
+
"learning_rate": 7.714384004615157e-06,
|
| 3000 |
+
"loss": 0.4187,
|
| 3001 |
+
"num_tokens": 1429272848.0,
|
| 3002 |
+
"step": 1870
|
| 3003 |
+
},
|
| 3004 |
+
{
|
| 3005 |
+
"epoch": 2.557980900409277,
|
| 3006 |
+
"grad_norm": 0.19850404598262095,
|
| 3007 |
+
"learning_rate": 7.634385231096583e-06,
|
| 3008 |
+
"loss": 0.4154,
|
| 3009 |
+
"num_tokens": 1433091944.0,
|
| 3010 |
+
"step": 1875
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"epoch": 2.5648021828103684,
|
| 3014 |
+
"grad_norm": 0.20192494347851303,
|
| 3015 |
+
"learning_rate": 7.555509672429645e-06,
|
| 3016 |
+
"loss": 0.4206,
|
| 3017 |
+
"num_tokens": 1436974840.0,
|
| 3018 |
+
"step": 1880
|
| 3019 |
+
},
|
| 3020 |
+
{
|
| 3021 |
+
"epoch": 2.57162346521146,
|
| 3022 |
+
"grad_norm": 0.2075190266205154,
|
| 3023 |
+
"learning_rate": 7.4777617882900305e-06,
|
| 3024 |
+
"loss": 0.4261,
|
| 3025 |
+
"num_tokens": 1440774914.0,
|
| 3026 |
+
"step": 1885
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 2.578444747612551,
|
| 3030 |
+
"grad_norm": 0.19747582047162396,
|
| 3031 |
+
"learning_rate": 7.401145974593946e-06,
|
| 3032 |
+
"loss": 0.4189,
|
| 3033 |
+
"num_tokens": 1444556279.0,
|
| 3034 |
+
"step": 1890
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 2.5852660300136425,
|
| 3038 |
+
"grad_norm": 0.21156834672027997,
|
| 3039 |
+
"learning_rate": 7.325666563249609e-06,
|
| 3040 |
+
"loss": 0.4242,
|
| 3041 |
+
"num_tokens": 1448438583.0,
|
| 3042 |
+
"step": 1895
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 2.592087312414734,
|
| 3046 |
+
"grad_norm": 0.19832796398348237,
|
| 3047 |
+
"learning_rate": 7.251327821912306e-06,
|
| 3048 |
+
"loss": 0.4282,
|
| 3049 |
+
"num_tokens": 1452265791.0,
|
| 3050 |
+
"step": 1900
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 2.5989085948158253,
|
| 3054 |
+
"grad_norm": 0.20969439420552538,
|
| 3055 |
+
"learning_rate": 7.178133953743086e-06,
|
| 3056 |
+
"loss": 0.4266,
|
| 3057 |
+
"num_tokens": 1455960099.0,
|
| 3058 |
+
"step": 1905
|
| 3059 |
+
},
|
| 3060 |
+
{
|
| 3061 |
+
"epoch": 2.6057298772169166,
|
| 3062 |
+
"grad_norm": 0.22844147412156932,
|
| 3063 |
+
"learning_rate": 7.10608909717113e-06,
|
| 3064 |
+
"loss": 0.4231,
|
| 3065 |
+
"num_tokens": 1459864902.0,
|
| 3066 |
+
"step": 1910
|
| 3067 |
+
},
|
| 3068 |
+
{
|
| 3069 |
+
"epoch": 2.6125511596180084,
|
| 3070 |
+
"grad_norm": 0.20300761017980706,
|
| 3071 |
+
"learning_rate": 7.035197325659751e-06,
|
| 3072 |
+
"loss": 0.412,
|
| 3073 |
+
"num_tokens": 1463560724.0,
|
| 3074 |
+
"step": 1915
|
| 3075 |
+
},
|
| 3076 |
+
{
|
| 3077 |
+
"epoch": 2.6193724420191,
|
| 3078 |
+
"grad_norm": 0.21799998044827612,
|
| 3079 |
+
"learning_rate": 6.965462647476069e-06,
|
| 3080 |
+
"loss": 0.4161,
|
| 3081 |
+
"num_tokens": 1467318533.0,
|
| 3082 |
+
"step": 1920
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 2.626193724420191,
|
| 3086 |
+
"grad_norm": 0.20530834564702366,
|
| 3087 |
+
"learning_rate": 6.896889005464401e-06,
|
| 3088 |
+
"loss": 0.4194,
|
| 3089 |
+
"num_tokens": 1471225951.0,
|
| 3090 |
+
"step": 1925
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 2.6330150068212825,
|
| 3094 |
+
"grad_norm": 0.23082948561064157,
|
| 3095 |
+
"learning_rate": 6.82948027682331e-06,
|
| 3096 |
+
"loss": 0.4149,
|
| 3097 |
+
"num_tokens": 1474976948.0,
|
| 3098 |
+
"step": 1930
|
| 3099 |
+
},
|
| 3100 |
+
{
|
| 3101 |
+
"epoch": 2.639836289222374,
|
| 3102 |
+
"grad_norm": 0.206898898979367,
|
| 3103 |
+
"learning_rate": 6.763240272886407e-06,
|
| 3104 |
+
"loss": 0.4188,
|
| 3105 |
+
"num_tokens": 1478830865.0,
|
| 3106 |
+
"step": 1935
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 2.6466575716234653,
|
| 3110 |
+
"grad_norm": 0.21323938935396114,
|
| 3111 |
+
"learning_rate": 6.698172738906834e-06,
|
| 3112 |
+
"loss": 0.4319,
|
| 3113 |
+
"num_tokens": 1482679267.0,
|
| 3114 |
+
"step": 1940
|
| 3115 |
+
},
|
| 3116 |
+
{
|
| 3117 |
+
"epoch": 2.6534788540245566,
|
| 3118 |
+
"grad_norm": 0.20405959106921281,
|
| 3119 |
+
"learning_rate": 6.634281353845513e-06,
|
| 3120 |
+
"loss": 0.4246,
|
| 3121 |
+
"num_tokens": 1486426058.0,
|
| 3122 |
+
"step": 1945
|
| 3123 |
+
},
|
| 3124 |
+
{
|
| 3125 |
+
"epoch": 2.660300136425648,
|
| 3126 |
+
"grad_norm": 0.19891088866150305,
|
| 3127 |
+
"learning_rate": 6.571569730163153e-06,
|
| 3128 |
+
"loss": 0.4111,
|
| 3129 |
+
"num_tokens": 1490166733.0,
|
| 3130 |
+
"step": 1950
|
| 3131 |
+
},
|
| 3132 |
+
{
|
| 3133 |
+
"epoch": 2.6671214188267394,
|
| 3134 |
+
"grad_norm": 0.2015858137358049,
|
| 3135 |
+
"learning_rate": 6.5100414136159595e-06,
|
| 3136 |
+
"loss": 0.4252,
|
| 3137 |
+
"num_tokens": 1493992930.0,
|
| 3138 |
+
"step": 1955
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 2.6739427012278307,
|
| 3142 |
+
"grad_norm": 0.20788919401571546,
|
| 3143 |
+
"learning_rate": 6.4496998830552045e-06,
|
| 3144 |
+
"loss": 0.4108,
|
| 3145 |
+
"num_tokens": 1497730531.0,
|
| 3146 |
+
"step": 1960
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 2.680763983628922,
|
| 3150 |
+
"grad_norm": 0.1942548104643525,
|
| 3151 |
+
"learning_rate": 6.390548550230478e-06,
|
| 3152 |
+
"loss": 0.4258,
|
| 3153 |
+
"num_tokens": 1501593876.0,
|
| 3154 |
+
"step": 1965
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.6875852660300135,
|
| 3158 |
+
"grad_norm": 0.21777864642203057,
|
| 3159 |
+
"learning_rate": 6.332590759596831e-06,
|
| 3160 |
+
"loss": 0.4225,
|
| 3161 |
+
"num_tokens": 1505411865.0,
|
| 3162 |
+
"step": 1970
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 2.694406548431105,
|
| 3166 |
+
"grad_norm": 0.1947562097326217,
|
| 3167 |
+
"learning_rate": 6.275829788125652e-06,
|
| 3168 |
+
"loss": 0.4154,
|
| 3169 |
+
"num_tokens": 1509039077.0,
|
| 3170 |
+
"step": 1975
|
| 3171 |
+
},
|
| 3172 |
+
{
|
| 3173 |
+
"epoch": 2.701227830832196,
|
| 3174 |
+
"grad_norm": 0.20152762212569178,
|
| 3175 |
+
"learning_rate": 6.220268845119388e-06,
|
| 3176 |
+
"loss": 0.4169,
|
| 3177 |
+
"num_tokens": 1512763400.0,
|
| 3178 |
+
"step": 1980
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 2.708049113233288,
|
| 3182 |
+
"grad_norm": 0.20050868114123066,
|
| 3183 |
+
"learning_rate": 6.1659110720301e-06,
|
| 3184 |
+
"loss": 0.4231,
|
| 3185 |
+
"num_tokens": 1516724445.0,
|
| 3186 |
+
"step": 1985
|
| 3187 |
+
},
|
| 3188 |
+
{
|
| 3189 |
+
"epoch": 2.7148703956343794,
|
| 3190 |
+
"grad_norm": 0.19329882456357955,
|
| 3191 |
+
"learning_rate": 6.112759542281827e-06,
|
| 3192 |
+
"loss": 0.416,
|
| 3193 |
+
"num_tokens": 1520614981.0,
|
| 3194 |
+
"step": 1990
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 2.7216916780354707,
|
| 3198 |
+
"grad_norm": 0.19853576242037266,
|
| 3199 |
+
"learning_rate": 6.0608172610968196e-06,
|
| 3200 |
+
"loss": 0.4238,
|
| 3201 |
+
"num_tokens": 1524397796.0,
|
| 3202 |
+
"step": 1995
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 2.728512960436562,
|
| 3206 |
+
"grad_norm": 0.19951479560606716,
|
| 3207 |
+
"learning_rate": 6.010087165325633e-06,
|
| 3208 |
+
"loss": 0.4189,
|
| 3209 |
+
"num_tokens": 1528158624.0,
|
| 3210 |
+
"step": 2000
|
| 3211 |
+
},
|
| 3212 |
+
{
|
| 3213 |
+
"epoch": 2.7353342428376535,
|
| 3214 |
+
"grad_norm": 0.20145252593989374,
|
| 3215 |
+
"learning_rate": 5.960572123281055e-06,
|
| 3216 |
+
"loss": 0.4323,
|
| 3217 |
+
"num_tokens": 1532032964.0,
|
| 3218 |
+
"step": 2005
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 2.742155525238745,
|
| 3222 |
+
"grad_norm": 0.20302873915041478,
|
| 3223 |
+
"learning_rate": 5.912274934575956e-06,
|
| 3224 |
+
"loss": 0.4214,
|
| 3225 |
+
"num_tokens": 1535931108.0,
|
| 3226 |
+
"step": 2010
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 2.748976807639836,
|
| 3230 |
+
"grad_norm": 0.19675090945354656,
|
| 3231 |
+
"learning_rate": 5.865198329964963e-06,
|
| 3232 |
+
"loss": 0.4088,
|
| 3233 |
+
"num_tokens": 1539611481.0,
|
| 3234 |
+
"step": 2015
|
| 3235 |
+
},
|
| 3236 |
+
{
|
| 3237 |
+
"epoch": 2.7557980900409276,
|
| 3238 |
+
"grad_norm": 0.20739294800070435,
|
| 3239 |
+
"learning_rate": 5.819344971190096e-06,
|
| 3240 |
+
"loss": 0.43,
|
| 3241 |
+
"num_tokens": 1543769258.0,
|
| 3242 |
+
"step": 2020
|
| 3243 |
+
},
|
| 3244 |
+
{
|
| 3245 |
+
"epoch": 2.762619372442019,
|
| 3246 |
+
"grad_norm": 0.19531202135351058,
|
| 3247 |
+
"learning_rate": 5.77471745083025e-06,
|
| 3248 |
+
"loss": 0.4217,
|
| 3249 |
+
"num_tokens": 1547704714.0,
|
| 3250 |
+
"step": 2025
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 2.7694406548431107,
|
| 3254 |
+
"grad_norm": 0.1933063488006224,
|
| 3255 |
+
"learning_rate": 5.731318292154611e-06,
|
| 3256 |
+
"loss": 0.4202,
|
| 3257 |
+
"num_tokens": 1551607979.0,
|
| 3258 |
+
"step": 2030
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 2.776261937244202,
|
| 3262 |
+
"grad_norm": 0.20842115092593327,
|
| 3263 |
+
"learning_rate": 5.689149948979993e-06,
|
| 3264 |
+
"loss": 0.4248,
|
| 3265 |
+
"num_tokens": 1555548130.0,
|
| 3266 |
+
"step": 2035
|
| 3267 |
+
},
|
| 3268 |
+
{
|
| 3269 |
+
"epoch": 2.7830832196452935,
|
| 3270 |
+
"grad_norm": 0.19224611133337063,
|
| 3271 |
+
"learning_rate": 5.648214805532103e-06,
|
| 3272 |
+
"loss": 0.4107,
|
| 3273 |
+
"num_tokens": 1559374280.0,
|
| 3274 |
+
"step": 2040
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 2.789904502046385,
|
| 3278 |
+
"grad_norm": 0.19419272835835372,
|
| 3279 |
+
"learning_rate": 5.6085151763107185e-06,
|
| 3280 |
+
"loss": 0.4226,
|
| 3281 |
+
"num_tokens": 1563231974.0,
|
| 3282 |
+
"step": 2045
|
| 3283 |
+
},
|
| 3284 |
+
{
|
| 3285 |
+
"epoch": 2.796725784447476,
|
| 3286 |
+
"grad_norm": 0.19848587760730213,
|
| 3287 |
+
"learning_rate": 5.570053305958847e-06,
|
| 3288 |
+
"loss": 0.428,
|
| 3289 |
+
"num_tokens": 1567090495.0,
|
| 3290 |
+
"step": 2050
|
| 3291 |
+
},
|
| 3292 |
+
{
|
| 3293 |
+
"epoch": 2.8035470668485676,
|
| 3294 |
+
"grad_norm": 0.20454851749822658,
|
| 3295 |
+
"learning_rate": 5.532831369135785e-06,
|
| 3296 |
+
"loss": 0.4326,
|
| 3297 |
+
"num_tokens": 1570887007.0,
|
| 3298 |
+
"step": 2055
|
| 3299 |
+
},
|
| 3300 |
+
{
|
| 3301 |
+
"epoch": 2.810368349249659,
|
| 3302 |
+
"grad_norm": 0.19640234313277483,
|
| 3303 |
+
"learning_rate": 5.496851470394195e-06,
|
| 3304 |
+
"loss": 0.4151,
|
| 3305 |
+
"num_tokens": 1574806787.0,
|
| 3306 |
+
"step": 2060
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 2.8171896316507503,
|
| 3310 |
+
"grad_norm": 0.19810517176834253,
|
| 3311 |
+
"learning_rate": 5.462115644061074e-06,
|
| 3312 |
+
"loss": 0.4205,
|
| 3313 |
+
"num_tokens": 1578732597.0,
|
| 3314 |
+
"step": 2065
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 2.8240109140518417,
|
| 3318 |
+
"grad_norm": 0.19768847828191524,
|
| 3319 |
+
"learning_rate": 5.428625854122761e-06,
|
| 3320 |
+
"loss": 0.4244,
|
| 3321 |
+
"num_tokens": 1582527990.0,
|
| 3322 |
+
"step": 2070
|
| 3323 |
+
},
|
| 3324 |
+
{
|
| 3325 |
+
"epoch": 2.830832196452933,
|
| 3326 |
+
"grad_norm": 0.19591138130670074,
|
| 3327 |
+
"learning_rate": 5.396383994113894e-06,
|
| 3328 |
+
"loss": 0.4123,
|
| 3329 |
+
"num_tokens": 1586163936.0,
|
| 3330 |
+
"step": 2075
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 2.8376534788540244,
|
| 3334 |
+
"grad_norm": 0.19970168035474128,
|
| 3335 |
+
"learning_rate": 5.365391887010311e-06,
|
| 3336 |
+
"loss": 0.4102,
|
| 3337 |
+
"num_tokens": 1589894282.0,
|
| 3338 |
+
"step": 2080
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"epoch": 2.844474761255116,
|
| 3342 |
+
"grad_norm": 0.20019629815450546,
|
| 3343 |
+
"learning_rate": 5.335651285126024e-06,
|
| 3344 |
+
"loss": 0.4255,
|
| 3345 |
+
"num_tokens": 1593669447.0,
|
| 3346 |
+
"step": 2085
|
| 3347 |
+
},
|
| 3348 |
+
{
|
| 3349 |
+
"epoch": 2.851296043656207,
|
| 3350 |
+
"grad_norm": 0.19850252932551976,
|
| 3351 |
+
"learning_rate": 5.307163870014117e-06,
|
| 3352 |
+
"loss": 0.4182,
|
| 3353 |
+
"num_tokens": 1597406094.0,
|
| 3354 |
+
"step": 2090
|
| 3355 |
+
},
|
| 3356 |
+
{
|
| 3357 |
+
"epoch": 2.8581173260572985,
|
| 3358 |
+
"grad_norm": 0.1844183685532561,
|
| 3359 |
+
"learning_rate": 5.279931252371664e-06,
|
| 3360 |
+
"loss": 0.4215,
|
| 3361 |
+
"num_tokens": 1601353966.0,
|
| 3362 |
+
"step": 2095
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 2.8649386084583903,
|
| 3366 |
+
"grad_norm": 0.20211606394427048,
|
| 3367 |
+
"learning_rate": 5.253954971948686e-06,
|
| 3368 |
+
"loss": 0.4256,
|
| 3369 |
+
"num_tokens": 1605354453.0,
|
| 3370 |
+
"step": 2100
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 2.8717598908594817,
|
| 3374 |
+
"grad_norm": 0.20345823217014605,
|
| 3375 |
+
"learning_rate": 5.229236497461059e-06,
|
| 3376 |
+
"loss": 0.4299,
|
| 3377 |
+
"num_tokens": 1609346678.0,
|
| 3378 |
+
"step": 2105
|
| 3379 |
+
},
|
| 3380 |
+
{
|
| 3381 |
+
"epoch": 2.878581173260573,
|
| 3382 |
+
"grad_norm": 0.1975895200056834,
|
| 3383 |
+
"learning_rate": 5.205777226507506e-06,
|
| 3384 |
+
"loss": 0.4251,
|
| 3385 |
+
"num_tokens": 1613213477.0,
|
| 3386 |
+
"step": 2110
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 2.8854024556616644,
|
| 3390 |
+
"grad_norm": 0.18996613901488008,
|
| 3391 |
+
"learning_rate": 5.183578485490543e-06,
|
| 3392 |
+
"loss": 0.4152,
|
| 3393 |
+
"num_tokens": 1616990063.0,
|
| 3394 |
+
"step": 2115
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 2.892223738062756,
|
| 3398 |
+
"grad_norm": 0.20664944883370504,
|
| 3399 |
+
"learning_rate": 5.162641529541507e-06,
|
| 3400 |
+
"loss": 0.4135,
|
| 3401 |
+
"num_tokens": 1620738563.0,
|
| 3402 |
+
"step": 2120
|
| 3403 |
+
},
|
| 3404 |
+
{
|
| 3405 |
+
"epoch": 2.899045020463847,
|
| 3406 |
+
"grad_norm": 0.20839715507501666,
|
| 3407 |
+
"learning_rate": 5.142967542449581e-06,
|
| 3408 |
+
"loss": 0.4273,
|
| 3409 |
+
"num_tokens": 1624503206.0,
|
| 3410 |
+
"step": 2125
|
| 3411 |
+
},
|
| 3412 |
+
{
|
| 3413 |
+
"epoch": 2.9058663028649385,
|
| 3414 |
+
"grad_norm": 0.19579863013472537,
|
| 3415 |
+
"learning_rate": 5.124557636594861e-06,
|
| 3416 |
+
"loss": 0.4238,
|
| 3417 |
+
"num_tokens": 1628389551.0,
|
| 3418 |
+
"step": 2130
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 2.91268758526603,
|
| 3422 |
+
"grad_norm": 0.20826204742316173,
|
| 3423 |
+
"learning_rate": 5.107412852885468e-06,
|
| 3424 |
+
"loss": 0.419,
|
| 3425 |
+
"num_tokens": 1632099673.0,
|
| 3426 |
+
"step": 2135
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 2.9195088676671213,
|
| 3430 |
+
"grad_norm": 0.18823980909615004,
|
| 3431 |
+
"learning_rate": 5.091534160698686e-06,
|
| 3432 |
+
"loss": 0.4229,
|
| 3433 |
+
"num_tokens": 1636056127.0,
|
| 3434 |
+
"step": 2140
|
| 3435 |
+
},
|
| 3436 |
+
{
|
| 3437 |
+
"epoch": 2.926330150068213,
|
| 3438 |
+
"grad_norm": 0.19965301204123226,
|
| 3439 |
+
"learning_rate": 5.076922457826151e-06,
|
| 3440 |
+
"loss": 0.4155,
|
| 3441 |
+
"num_tokens": 1639823774.0,
|
| 3442 |
+
"step": 2145
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 2.9331514324693044,
|
| 3446 |
+
"grad_norm": 0.19783373733630943,
|
| 3447 |
+
"learning_rate": 5.063578570423107e-06,
|
| 3448 |
+
"loss": 0.4228,
|
| 3449 |
+
"num_tokens": 1643643243.0,
|
| 3450 |
+
"step": 2150
|
| 3451 |
+
},
|
| 3452 |
+
{
|
| 3453 |
+
"epoch": 2.939972714870396,
|
| 3454 |
+
"grad_norm": 0.1918627268234313,
|
| 3455 |
+
"learning_rate": 5.051503252961668e-06,
|
| 3456 |
+
"loss": 0.4157,
|
| 3457 |
+
"num_tokens": 1647326744.0,
|
| 3458 |
+
"step": 2155
|
| 3459 |
+
},
|
| 3460 |
+
{
|
| 3461 |
+
"epoch": 2.946793997271487,
|
| 3462 |
+
"grad_norm": 0.19853326107825944,
|
| 3463 |
+
"learning_rate": 5.040697188188186e-06,
|
| 3464 |
+
"loss": 0.4148,
|
| 3465 |
+
"num_tokens": 1651209275.0,
|
| 3466 |
+
"step": 2160
|
| 3467 |
+
},
|
| 3468 |
+
{
|
| 3469 |
+
"epoch": 2.9536152796725785,
|
| 3470 |
+
"grad_norm": 0.21266069031952056,
|
| 3471 |
+
"learning_rate": 5.031160987084629e-06,
|
| 3472 |
+
"loss": 0.4219,
|
| 3473 |
+
"num_tokens": 1654945680.0,
|
| 3474 |
+
"step": 2165
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 2.96043656207367,
|
| 3478 |
+
"grad_norm": 0.19519354152848462,
|
| 3479 |
+
"learning_rate": 5.022895188834045e-06,
|
| 3480 |
+
"loss": 0.4148,
|
| 3481 |
+
"num_tokens": 1658670255.0,
|
| 3482 |
+
"step": 2170
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 2.9672578444747613,
|
| 3486 |
+
"grad_norm": 0.20169693136976627,
|
| 3487 |
+
"learning_rate": 5.015900260790069e-06,
|
| 3488 |
+
"loss": 0.4234,
|
| 3489 |
+
"num_tokens": 1662421467.0,
|
| 3490 |
+
"step": 2175
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"epoch": 2.9740791268758526,
|
| 3494 |
+
"grad_norm": 0.19592307614732052,
|
| 3495 |
+
"learning_rate": 5.010176598450514e-06,
|
| 3496 |
+
"loss": 0.4179,
|
| 3497 |
+
"num_tokens": 1666233128.0,
|
| 3498 |
+
"step": 2180
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 2.980900409276944,
|
| 3502 |
+
"grad_norm": 0.20871410874190424,
|
| 3503 |
+
"learning_rate": 5.005724525434989e-06,
|
| 3504 |
+
"loss": 0.423,
|
| 3505 |
+
"num_tokens": 1670143796.0,
|
| 3506 |
+
"step": 2185
|
| 3507 |
+
},
|
| 3508 |
+
{
|
| 3509 |
+
"epoch": 2.9877216916780354,
|
| 3510 |
+
"grad_norm": 0.19352266138909222,
|
| 3511 |
+
"learning_rate": 5.002544293466611e-06,
|
| 3512 |
+
"loss": 0.4189,
|
| 3513 |
+
"num_tokens": 1674122886.0,
|
| 3514 |
+
"step": 2190
|
| 3515 |
+
},
|
| 3516 |
+
{
|
| 3517 |
+
"epoch": 2.9945429740791267,
|
| 3518 |
+
"grad_norm": 0.1994426341020866,
|
| 3519 |
+
"learning_rate": 5.000636082357781e-06,
|
| 3520 |
+
"loss": 0.4213,
|
| 3521 |
+
"num_tokens": 1677987646.0,
|
| 3522 |
+
"step": 2195
|
| 3523 |
+
},
|
| 3524 |
+
{
|
| 3525 |
+
"epoch": 3.0,
|
| 3526 |
+
"num_tokens": 1680919540.0,
|
| 3527 |
+
"step": 2199,
|
| 3528 |
+
"total_flos": 3438830624964608.0,
|
| 3529 |
+
"train_loss": 0.4748283546585666,
|
| 3530 |
+
"train_runtime": 17138.2816,
|
| 3531 |
+
"train_samples_per_second": 16.408,
|
| 3532 |
+
"train_steps_per_second": 0.128
|
| 3533 |
+
}
|
| 3534 |
+
],
|
| 3535 |
+
"logging_steps": 5,
|
| 3536 |
+
"max_steps": 2199,
|
| 3537 |
+
"num_input_tokens_seen": 0,
|
| 3538 |
+
"num_train_epochs": 3,
|
| 3539 |
+
"save_steps": 100,
|
| 3540 |
+
"stateful_callbacks": {
|
| 3541 |
+
"TrainerControl": {
|
| 3542 |
+
"args": {
|
| 3543 |
+
"should_epoch_stop": false,
|
| 3544 |
+
"should_evaluate": false,
|
| 3545 |
+
"should_log": false,
|
| 3546 |
+
"should_save": true,
|
| 3547 |
+
"should_training_stop": true
|
| 3548 |
+
},
|
| 3549 |
+
"attributes": {}
|
| 3550 |
+
}
|
| 3551 |
+
},
|
| 3552 |
+
"total_flos": 3438830624964608.0,
|
| 3553 |
+
"train_batch_size": 16,
|
| 3554 |
+
"trial_name": null,
|
| 3555 |
+
"trial_params": null
|
| 3556 |
+
}
|