Model save
Browse files- README.md +57 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +3088 -0
README.md
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: mistralai/Mistral-7B-v0.3
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Mistral-7B-v0.3-sft-ultrachat
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- sft
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Mistral-7B-v0.3-sft-ultrachat
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.3](https://huggingface.co/mistralai/Mistral-7B-v0.3).
|
| 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="AmberYifan/Mistral-7B-v0.3-sft-ultrachat", 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 |
+
|
| 31 |
+
|
| 32 |
+
This model was trained with SFT.
|
| 33 |
+
|
| 34 |
+
### Framework versions
|
| 35 |
+
|
| 36 |
+
- TRL: 0.12.2
|
| 37 |
+
- Transformers: 4.46.3
|
| 38 |
+
- Pytorch: 2.5.1+cu118
|
| 39 |
+
- Datasets: 3.2.0
|
| 40 |
+
- Tokenizers: 0.20.3
|
| 41 |
+
|
| 42 |
+
## Citations
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
Cite TRL as:
|
| 47 |
+
|
| 48 |
+
```bibtex
|
| 49 |
+
@misc{vonwerra2022trl,
|
| 50 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 51 |
+
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 Gallouédec},
|
| 52 |
+
year = 2020,
|
| 53 |
+
journal = {GitHub repository},
|
| 54 |
+
publisher = {GitHub},
|
| 55 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 56 |
+
}
|
| 57 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 113379083550720.0,
|
| 4 |
+
"train_loss": 1.171416565762112,
|
| 5 |
+
"train_runtime": 11018.7418,
|
| 6 |
+
"train_samples": 51966,
|
| 7 |
+
"train_samples_per_second": 6.29,
|
| 8 |
+
"train_steps_per_second": 0.197
|
| 9 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"transformers_version": "4.46.3"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"total_flos": 113379083550720.0,
|
| 4 |
+
"train_loss": 1.171416565762112,
|
| 5 |
+
"train_runtime": 11018.7418,
|
| 6 |
+
"train_samples": 51966,
|
| 7 |
+
"train_samples_per_second": 6.29,
|
| 8 |
+
"train_steps_per_second": 0.197
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3088 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2166,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0004616805170821791,
|
| 13 |
+
"grad_norm": 15.672144611335952,
|
| 14 |
+
"learning_rate": 9.216589861751152e-08,
|
| 15 |
+
"loss": 1.3168,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0023084025854108957,
|
| 20 |
+
"grad_norm": 14.680930201512727,
|
| 21 |
+
"learning_rate": 4.608294930875577e-07,
|
| 22 |
+
"loss": 1.2513,
|
| 23 |
+
"step": 5
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0046168051708217915,
|
| 27 |
+
"grad_norm": 8.024031806354767,
|
| 28 |
+
"learning_rate": 9.216589861751154e-07,
|
| 29 |
+
"loss": 1.1965,
|
| 30 |
+
"step": 10
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.006925207756232687,
|
| 34 |
+
"grad_norm": 5.48716759630272,
|
| 35 |
+
"learning_rate": 1.382488479262673e-06,
|
| 36 |
+
"loss": 1.0985,
|
| 37 |
+
"step": 15
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.009233610341643583,
|
| 41 |
+
"grad_norm": 5.068857847099135,
|
| 42 |
+
"learning_rate": 1.8433179723502307e-06,
|
| 43 |
+
"loss": 1.0416,
|
| 44 |
+
"step": 20
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.011542012927054479,
|
| 48 |
+
"grad_norm": 5.784416390233076,
|
| 49 |
+
"learning_rate": 2.3041474654377884e-06,
|
| 50 |
+
"loss": 1.0414,
|
| 51 |
+
"step": 25
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.013850415512465374,
|
| 55 |
+
"grad_norm": 5.088059220204017,
|
| 56 |
+
"learning_rate": 2.764976958525346e-06,
|
| 57 |
+
"loss": 1.0977,
|
| 58 |
+
"step": 30
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.016158818097876268,
|
| 62 |
+
"grad_norm": 4.831013532516624,
|
| 63 |
+
"learning_rate": 3.225806451612903e-06,
|
| 64 |
+
"loss": 1.0864,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.018467220683287166,
|
| 69 |
+
"grad_norm": 5.01522341202905,
|
| 70 |
+
"learning_rate": 3.6866359447004615e-06,
|
| 71 |
+
"loss": 1.0875,
|
| 72 |
+
"step": 40
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02077562326869806,
|
| 76 |
+
"grad_norm": 5.2743430965797495,
|
| 77 |
+
"learning_rate": 4.147465437788019e-06,
|
| 78 |
+
"loss": 1.0582,
|
| 79 |
+
"step": 45
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.023084025854108958,
|
| 83 |
+
"grad_norm": 4.868797420697276,
|
| 84 |
+
"learning_rate": 4.608294930875577e-06,
|
| 85 |
+
"loss": 1.0706,
|
| 86 |
+
"step": 50
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.025392428439519853,
|
| 90 |
+
"grad_norm": 4.866881384682284,
|
| 91 |
+
"learning_rate": 5.0691244239631346e-06,
|
| 92 |
+
"loss": 1.0874,
|
| 93 |
+
"step": 55
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.027700831024930747,
|
| 97 |
+
"grad_norm": 4.942215298185941,
|
| 98 |
+
"learning_rate": 5.529953917050692e-06,
|
| 99 |
+
"loss": 1.0765,
|
| 100 |
+
"step": 60
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.030009233610341645,
|
| 104 |
+
"grad_norm": 4.510624758005134,
|
| 105 |
+
"learning_rate": 5.9907834101382485e-06,
|
| 106 |
+
"loss": 1.0651,
|
| 107 |
+
"step": 65
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.032317636195752536,
|
| 111 |
+
"grad_norm": 4.961160516804359,
|
| 112 |
+
"learning_rate": 6.451612903225806e-06,
|
| 113 |
+
"loss": 1.0826,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03462603878116344,
|
| 118 |
+
"grad_norm": 4.999318871040395,
|
| 119 |
+
"learning_rate": 6.912442396313365e-06,
|
| 120 |
+
"loss": 1.0966,
|
| 121 |
+
"step": 75
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.03693444136657433,
|
| 125 |
+
"grad_norm": 4.639315170945839,
|
| 126 |
+
"learning_rate": 7.373271889400923e-06,
|
| 127 |
+
"loss": 1.0934,
|
| 128 |
+
"step": 80
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.039242843951985226,
|
| 132 |
+
"grad_norm": 4.79620290699333,
|
| 133 |
+
"learning_rate": 7.83410138248848e-06,
|
| 134 |
+
"loss": 1.0932,
|
| 135 |
+
"step": 85
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04155124653739612,
|
| 139 |
+
"grad_norm": 4.957993386602933,
|
| 140 |
+
"learning_rate": 8.294930875576038e-06,
|
| 141 |
+
"loss": 1.1032,
|
| 142 |
+
"step": 90
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.043859649122807015,
|
| 146 |
+
"grad_norm": 4.669607524515842,
|
| 147 |
+
"learning_rate": 8.755760368663595e-06,
|
| 148 |
+
"loss": 1.0875,
|
| 149 |
+
"step": 95
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.046168051708217916,
|
| 153 |
+
"grad_norm": 4.602861109332021,
|
| 154 |
+
"learning_rate": 9.216589861751153e-06,
|
| 155 |
+
"loss": 1.0809,
|
| 156 |
+
"step": 100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.04847645429362881,
|
| 160 |
+
"grad_norm": 4.548726172146522,
|
| 161 |
+
"learning_rate": 9.67741935483871e-06,
|
| 162 |
+
"loss": 1.1218,
|
| 163 |
+
"step": 105
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.050784856879039705,
|
| 167 |
+
"grad_norm": 4.724149571099335,
|
| 168 |
+
"learning_rate": 1.0138248847926269e-05,
|
| 169 |
+
"loss": 1.1007,
|
| 170 |
+
"step": 110
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.0530932594644506,
|
| 174 |
+
"grad_norm": 5.309010432204349,
|
| 175 |
+
"learning_rate": 1.0599078341013826e-05,
|
| 176 |
+
"loss": 1.1368,
|
| 177 |
+
"step": 115
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.055401662049861494,
|
| 181 |
+
"grad_norm": 4.839305137997795,
|
| 182 |
+
"learning_rate": 1.1059907834101385e-05,
|
| 183 |
+
"loss": 1.1055,
|
| 184 |
+
"step": 120
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05771006463527239,
|
| 188 |
+
"grad_norm": 4.5796294161615565,
|
| 189 |
+
"learning_rate": 1.152073732718894e-05,
|
| 190 |
+
"loss": 1.1155,
|
| 191 |
+
"step": 125
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06001846722068329,
|
| 195 |
+
"grad_norm": 4.706959812240538,
|
| 196 |
+
"learning_rate": 1.1981566820276497e-05,
|
| 197 |
+
"loss": 1.1387,
|
| 198 |
+
"step": 130
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.062326869806094184,
|
| 202 |
+
"grad_norm": 4.492165983938348,
|
| 203 |
+
"learning_rate": 1.2442396313364056e-05,
|
| 204 |
+
"loss": 1.1733,
|
| 205 |
+
"step": 135
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06463527239150507,
|
| 209 |
+
"grad_norm": 4.746032213098828,
|
| 210 |
+
"learning_rate": 1.2903225806451613e-05,
|
| 211 |
+
"loss": 1.1375,
|
| 212 |
+
"step": 140
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06694367497691597,
|
| 216 |
+
"grad_norm": 4.713817907356248,
|
| 217 |
+
"learning_rate": 1.3364055299539171e-05,
|
| 218 |
+
"loss": 1.158,
|
| 219 |
+
"step": 145
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06925207756232687,
|
| 223 |
+
"grad_norm": 4.342905646572964,
|
| 224 |
+
"learning_rate": 1.382488479262673e-05,
|
| 225 |
+
"loss": 1.1607,
|
| 226 |
+
"step": 150
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07156048014773776,
|
| 230 |
+
"grad_norm": 4.502102336400582,
|
| 231 |
+
"learning_rate": 1.4285714285714287e-05,
|
| 232 |
+
"loss": 1.1382,
|
| 233 |
+
"step": 155
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07386888273314866,
|
| 237 |
+
"grad_norm": 4.300393542300411,
|
| 238 |
+
"learning_rate": 1.4746543778801846e-05,
|
| 239 |
+
"loss": 1.1518,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.07617728531855955,
|
| 244 |
+
"grad_norm": 4.400546990325483,
|
| 245 |
+
"learning_rate": 1.5207373271889403e-05,
|
| 246 |
+
"loss": 1.1436,
|
| 247 |
+
"step": 165
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07848568790397045,
|
| 251 |
+
"grad_norm": 4.77590791643038,
|
| 252 |
+
"learning_rate": 1.566820276497696e-05,
|
| 253 |
+
"loss": 1.2173,
|
| 254 |
+
"step": 170
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08079409048938135,
|
| 258 |
+
"grad_norm": 4.32969974114785,
|
| 259 |
+
"learning_rate": 1.6129032258064517e-05,
|
| 260 |
+
"loss": 1.1654,
|
| 261 |
+
"step": 175
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08310249307479224,
|
| 265 |
+
"grad_norm": 5.285074448558262,
|
| 266 |
+
"learning_rate": 1.6589861751152075e-05,
|
| 267 |
+
"loss": 1.2185,
|
| 268 |
+
"step": 180
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08541089566020314,
|
| 272 |
+
"grad_norm": 6.312179413881035,
|
| 273 |
+
"learning_rate": 1.705069124423963e-05,
|
| 274 |
+
"loss": 1.2063,
|
| 275 |
+
"step": 185
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.08771929824561403,
|
| 279 |
+
"grad_norm": 4.351482667809684,
|
| 280 |
+
"learning_rate": 1.751152073732719e-05,
|
| 281 |
+
"loss": 1.1814,
|
| 282 |
+
"step": 190
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09002770083102493,
|
| 286 |
+
"grad_norm": 4.468079454686115,
|
| 287 |
+
"learning_rate": 1.7972350230414748e-05,
|
| 288 |
+
"loss": 1.2058,
|
| 289 |
+
"step": 195
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09233610341643583,
|
| 293 |
+
"grad_norm": 5.51425273025908,
|
| 294 |
+
"learning_rate": 1.8433179723502307e-05,
|
| 295 |
+
"loss": 1.1646,
|
| 296 |
+
"step": 200
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09464450600184672,
|
| 300 |
+
"grad_norm": 4.661323669253999,
|
| 301 |
+
"learning_rate": 1.8894009216589862e-05,
|
| 302 |
+
"loss": 1.1689,
|
| 303 |
+
"step": 205
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.09695290858725762,
|
| 307 |
+
"grad_norm": 726.888849011745,
|
| 308 |
+
"learning_rate": 1.935483870967742e-05,
|
| 309 |
+
"loss": 1.8785,
|
| 310 |
+
"step": 210
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.09926131117266851,
|
| 314 |
+
"grad_norm": 5.867844131835615,
|
| 315 |
+
"learning_rate": 1.981566820276498e-05,
|
| 316 |
+
"loss": 1.2661,
|
| 317 |
+
"step": 215
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.10156971375807941,
|
| 321 |
+
"grad_norm": 6.2684199277472015,
|
| 322 |
+
"learning_rate": 1.9999883080288618e-05,
|
| 323 |
+
"loss": 1.2545,
|
| 324 |
+
"step": 220
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.1038781163434903,
|
| 328 |
+
"grad_norm": 5.426004523317811,
|
| 329 |
+
"learning_rate": 1.999916858084231e-05,
|
| 330 |
+
"loss": 1.2259,
|
| 331 |
+
"step": 225
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1061865189289012,
|
| 335 |
+
"grad_norm": 4.617593739028291,
|
| 336 |
+
"learning_rate": 1.999780458369908e-05,
|
| 337 |
+
"loss": 1.177,
|
| 338 |
+
"step": 230
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.1084949215143121,
|
| 342 |
+
"grad_norm": 4.412649452769939,
|
| 343 |
+
"learning_rate": 1.9995791177457598e-05,
|
| 344 |
+
"loss": 1.2127,
|
| 345 |
+
"step": 235
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.11080332409972299,
|
| 349 |
+
"grad_norm": 4.3422685444059965,
|
| 350 |
+
"learning_rate": 1.9993128492899012e-05,
|
| 351 |
+
"loss": 1.2398,
|
| 352 |
+
"step": 240
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.11311172668513389,
|
| 356 |
+
"grad_norm": 4.837187367612426,
|
| 357 |
+
"learning_rate": 1.9989816702978447e-05,
|
| 358 |
+
"loss": 1.2189,
|
| 359 |
+
"step": 245
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11542012927054478,
|
| 363 |
+
"grad_norm": 4.203624655101154,
|
| 364 |
+
"learning_rate": 1.998585602281378e-05,
|
| 365 |
+
"loss": 1.1641,
|
| 366 |
+
"step": 250
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.11772853185595568,
|
| 370 |
+
"grad_norm": 4.172078683953242,
|
| 371 |
+
"learning_rate": 1.9981246709671668e-05,
|
| 372 |
+
"loss": 1.217,
|
| 373 |
+
"step": 255
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.12003693444136658,
|
| 377 |
+
"grad_norm": 4.445815868978626,
|
| 378 |
+
"learning_rate": 1.9975989062950828e-05,
|
| 379 |
+
"loss": 1.2198,
|
| 380 |
+
"step": 260
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.12234533702677747,
|
| 384 |
+
"grad_norm": 4.5591861583880045,
|
| 385 |
+
"learning_rate": 1.9970083424162598e-05,
|
| 386 |
+
"loss": 1.2971,
|
| 387 |
+
"step": 265
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12465373961218837,
|
| 391 |
+
"grad_norm": 8.794456155689286,
|
| 392 |
+
"learning_rate": 1.9963530176908752e-05,
|
| 393 |
+
"loss": 1.2543,
|
| 394 |
+
"step": 270
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.12696214219759927,
|
| 398 |
+
"grad_norm": 4.296337355363852,
|
| 399 |
+
"learning_rate": 1.9956329746856583e-05,
|
| 400 |
+
"loss": 1.1902,
|
| 401 |
+
"step": 275
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.12927054478301014,
|
| 405 |
+
"grad_norm": 4.210037276606183,
|
| 406 |
+
"learning_rate": 1.9948482601711245e-05,
|
| 407 |
+
"loss": 1.2119,
|
| 408 |
+
"step": 280
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13157894736842105,
|
| 412 |
+
"grad_norm": 4.525819047133829,
|
| 413 |
+
"learning_rate": 1.9939989251185386e-05,
|
| 414 |
+
"loss": 1.2267,
|
| 415 |
+
"step": 285
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.13388734995383195,
|
| 419 |
+
"grad_norm": 4.617114491344834,
|
| 420 |
+
"learning_rate": 1.993085024696604e-05,
|
| 421 |
+
"loss": 1.253,
|
| 422 |
+
"step": 290
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.13619575253924285,
|
| 426 |
+
"grad_norm": 4.306884827311129,
|
| 427 |
+
"learning_rate": 1.992106618267878e-05,
|
| 428 |
+
"loss": 1.2968,
|
| 429 |
+
"step": 295
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.13850415512465375,
|
| 433 |
+
"grad_norm": 4.06156778374873,
|
| 434 |
+
"learning_rate": 1.9910637693849166e-05,
|
| 435 |
+
"loss": 1.2523,
|
| 436 |
+
"step": 300
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.14081255771006462,
|
| 440 |
+
"grad_norm": 4.1725194301873225,
|
| 441 |
+
"learning_rate": 1.9899565457861463e-05,
|
| 442 |
+
"loss": 1.2465,
|
| 443 |
+
"step": 305
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.14312096029547552,
|
| 447 |
+
"grad_norm": 6.1582094432239005,
|
| 448 |
+
"learning_rate": 1.988785019391465e-05,
|
| 449 |
+
"loss": 1.2893,
|
| 450 |
+
"step": 310
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.14542936288088643,
|
| 454 |
+
"grad_norm": 4.468419480755536,
|
| 455 |
+
"learning_rate": 1.987549266297568e-05,
|
| 456 |
+
"loss": 1.2684,
|
| 457 |
+
"step": 315
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.14773776546629733,
|
| 461 |
+
"grad_norm": 4.663314719431853,
|
| 462 |
+
"learning_rate": 1.986249366773009e-05,
|
| 463 |
+
"loss": 1.2472,
|
| 464 |
+
"step": 320
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15004616805170823,
|
| 468 |
+
"grad_norm": 4.557295583444763,
|
| 469 |
+
"learning_rate": 1.9848854052529822e-05,
|
| 470 |
+
"loss": 1.2856,
|
| 471 |
+
"step": 325
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.1523545706371191,
|
| 475 |
+
"grad_norm": 4.128322557091226,
|
| 476 |
+
"learning_rate": 1.9834574703338406e-05,
|
| 477 |
+
"loss": 1.2717,
|
| 478 |
+
"step": 330
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.15466297322253,
|
| 482 |
+
"grad_norm": 4.265562249971871,
|
| 483 |
+
"learning_rate": 1.9819656547673393e-05,
|
| 484 |
+
"loss": 1.2614,
|
| 485 |
+
"step": 335
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.1569713758079409,
|
| 489 |
+
"grad_norm": 4.189648461283852,
|
| 490 |
+
"learning_rate": 1.9804100554546127e-05,
|
| 491 |
+
"loss": 1.2221,
|
| 492 |
+
"step": 340
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.1592797783933518,
|
| 496 |
+
"grad_norm": 4.753781028146877,
|
| 497 |
+
"learning_rate": 1.9787907734398785e-05,
|
| 498 |
+
"loss": 1.2641,
|
| 499 |
+
"step": 345
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.1615881809787627,
|
| 503 |
+
"grad_norm": 4.624571477954883,
|
| 504 |
+
"learning_rate": 1.9771079139038765e-05,
|
| 505 |
+
"loss": 1.3082,
|
| 506 |
+
"step": 350
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.16389658356417358,
|
| 510 |
+
"grad_norm": 4.51192227446534,
|
| 511 |
+
"learning_rate": 1.9753615861570338e-05,
|
| 512 |
+
"loss": 1.3116,
|
| 513 |
+
"step": 355
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.16620498614958448,
|
| 517 |
+
"grad_norm": 4.392313959090253,
|
| 518 |
+
"learning_rate": 1.9735519036323656e-05,
|
| 519 |
+
"loss": 1.2304,
|
| 520 |
+
"step": 360
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.16851338873499538,
|
| 524 |
+
"grad_norm": 4.979240339208881,
|
| 525 |
+
"learning_rate": 1.9716789838781095e-05,
|
| 526 |
+
"loss": 1.2682,
|
| 527 |
+
"step": 365
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.17082179132040629,
|
| 531 |
+
"grad_norm": 4.96937836441046,
|
| 532 |
+
"learning_rate": 1.9697429485500862e-05,
|
| 533 |
+
"loss": 1.3054,
|
| 534 |
+
"step": 370
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.1731301939058172,
|
| 538 |
+
"grad_norm": 3.935739346153204,
|
| 539 |
+
"learning_rate": 1.9677439234038004e-05,
|
| 540 |
+
"loss": 1.2704,
|
| 541 |
+
"step": 375
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.17543859649122806,
|
| 545 |
+
"grad_norm": 4.366123456450803,
|
| 546 |
+
"learning_rate": 1.96568203828627e-05,
|
| 547 |
+
"loss": 1.236,
|
| 548 |
+
"step": 380
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.17774699907663896,
|
| 552 |
+
"grad_norm": 4.003638705307624,
|
| 553 |
+
"learning_rate": 1.963557427127594e-05,
|
| 554 |
+
"loss": 1.2134,
|
| 555 |
+
"step": 385
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.18005540166204986,
|
| 559 |
+
"grad_norm": 4.711836278485082,
|
| 560 |
+
"learning_rate": 1.9613702279322518e-05,
|
| 561 |
+
"loss": 1.2424,
|
| 562 |
+
"step": 390
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.18236380424746076,
|
| 566 |
+
"grad_norm": 4.7756346414851345,
|
| 567 |
+
"learning_rate": 1.95912058277014e-05,
|
| 568 |
+
"loss": 1.2513,
|
| 569 |
+
"step": 395
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.18467220683287167,
|
| 573 |
+
"grad_norm": 4.055556447653374,
|
| 574 |
+
"learning_rate": 1.9568086377673422e-05,
|
| 575 |
+
"loss": 1.2305,
|
| 576 |
+
"step": 400
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.18698060941828254,
|
| 580 |
+
"grad_norm": 3.9870929086001605,
|
| 581 |
+
"learning_rate": 1.9544345430966398e-05,
|
| 582 |
+
"loss": 1.2766,
|
| 583 |
+
"step": 405
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.18928901200369344,
|
| 587 |
+
"grad_norm": 4.3683569271591525,
|
| 588 |
+
"learning_rate": 1.951998452967756e-05,
|
| 589 |
+
"loss": 1.2701,
|
| 590 |
+
"step": 410
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.19159741458910434,
|
| 594 |
+
"grad_norm": 4.282177503327308,
|
| 595 |
+
"learning_rate": 1.9495005256173398e-05,
|
| 596 |
+
"loss": 1.2173,
|
| 597 |
+
"step": 415
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.19390581717451524,
|
| 601 |
+
"grad_norm": 4.122228465513596,
|
| 602 |
+
"learning_rate": 1.9469409232986876e-05,
|
| 603 |
+
"loss": 1.293,
|
| 604 |
+
"step": 420
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.19621421975992612,
|
| 608 |
+
"grad_norm": 4.391730062186428,
|
| 609 |
+
"learning_rate": 1.9443198122712036e-05,
|
| 610 |
+
"loss": 1.3013,
|
| 611 |
+
"step": 425
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.19852262234533702,
|
| 615 |
+
"grad_norm": 4.2533205751093,
|
| 616 |
+
"learning_rate": 1.9416373627896002e-05,
|
| 617 |
+
"loss": 1.2478,
|
| 618 |
+
"step": 430
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.20083102493074792,
|
| 622 |
+
"grad_norm": 4.982151398275928,
|
| 623 |
+
"learning_rate": 1.9388937490928402e-05,
|
| 624 |
+
"loss": 1.289,
|
| 625 |
+
"step": 435
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.20313942751615882,
|
| 629 |
+
"grad_norm": 4.254393940238592,
|
| 630 |
+
"learning_rate": 1.9360891493928186e-05,
|
| 631 |
+
"loss": 1.2773,
|
| 632 |
+
"step": 440
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.20544783010156972,
|
| 636 |
+
"grad_norm": 4.812233488623846,
|
| 637 |
+
"learning_rate": 1.933223745862786e-05,
|
| 638 |
+
"loss": 1.2571,
|
| 639 |
+
"step": 445
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.2077562326869806,
|
| 643 |
+
"grad_norm": 4.193819364681046,
|
| 644 |
+
"learning_rate": 1.930297724625516e-05,
|
| 645 |
+
"loss": 1.3167,
|
| 646 |
+
"step": 450
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.2100646352723915,
|
| 650 |
+
"grad_norm": 4.318967687699199,
|
| 651 |
+
"learning_rate": 1.9273112757412165e-05,
|
| 652 |
+
"loss": 1.2578,
|
| 653 |
+
"step": 455
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.2123730378578024,
|
| 657 |
+
"grad_norm": 4.021438837096732,
|
| 658 |
+
"learning_rate": 1.9242645931951833e-05,
|
| 659 |
+
"loss": 1.2703,
|
| 660 |
+
"step": 460
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.2146814404432133,
|
| 664 |
+
"grad_norm": 3.9988355301981344,
|
| 665 |
+
"learning_rate": 1.921157874885199e-05,
|
| 666 |
+
"loss": 1.2702,
|
| 667 |
+
"step": 465
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.2169898430286242,
|
| 671 |
+
"grad_norm": 3.866018897785007,
|
| 672 |
+
"learning_rate": 1.91799132260868e-05,
|
| 673 |
+
"loss": 1.2651,
|
| 674 |
+
"step": 470
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.21929824561403508,
|
| 678 |
+
"grad_norm": 4.228145732575894,
|
| 679 |
+
"learning_rate": 1.9147651420495696e-05,
|
| 680 |
+
"loss": 1.2429,
|
| 681 |
+
"step": 475
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.22160664819944598,
|
| 685 |
+
"grad_norm": 4.16044625111994,
|
| 686 |
+
"learning_rate": 1.9114795427649735e-05,
|
| 687 |
+
"loss": 1.2263,
|
| 688 |
+
"step": 480
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.22391505078485688,
|
| 692 |
+
"grad_norm": 3.7071606709047678,
|
| 693 |
+
"learning_rate": 1.9081347381715535e-05,
|
| 694 |
+
"loss": 1.2592,
|
| 695 |
+
"step": 485
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.22622345337026778,
|
| 699 |
+
"grad_norm": 4.093983584879632,
|
| 700 |
+
"learning_rate": 1.904730945531661e-05,
|
| 701 |
+
"loss": 1.2819,
|
| 702 |
+
"step": 490
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.22853185595567868,
|
| 706 |
+
"grad_norm": 4.247421291613911,
|
| 707 |
+
"learning_rate": 1.901268385939226e-05,
|
| 708 |
+
"loss": 1.3118,
|
| 709 |
+
"step": 495
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.23084025854108955,
|
| 713 |
+
"grad_norm": 4.088704419142061,
|
| 714 |
+
"learning_rate": 1.8977472843053962e-05,
|
| 715 |
+
"loss": 1.2529,
|
| 716 |
+
"step": 500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.23314866112650046,
|
| 720 |
+
"grad_norm": 3.9526614218286698,
|
| 721 |
+
"learning_rate": 1.8941678693439272e-05,
|
| 722 |
+
"loss": 1.2254,
|
| 723 |
+
"step": 505
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.23545706371191136,
|
| 727 |
+
"grad_norm": 3.767319095108075,
|
| 728 |
+
"learning_rate": 1.8905303735563274e-05,
|
| 729 |
+
"loss": 1.2705,
|
| 730 |
+
"step": 510
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.23776546629732226,
|
| 734 |
+
"grad_norm": 4.1464464034097,
|
| 735 |
+
"learning_rate": 1.886835033216755e-05,
|
| 736 |
+
"loss": 1.2841,
|
| 737 |
+
"step": 515
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.24007386888273316,
|
| 741 |
+
"grad_norm": 4.154511161776497,
|
| 742 |
+
"learning_rate": 1.88308208835667e-05,
|
| 743 |
+
"loss": 1.2715,
|
| 744 |
+
"step": 520
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.24238227146814403,
|
| 748 |
+
"grad_norm": 4.815166096996458,
|
| 749 |
+
"learning_rate": 1.8792717827492446e-05,
|
| 750 |
+
"loss": 1.3034,
|
| 751 |
+
"step": 525
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.24469067405355494,
|
| 755 |
+
"grad_norm": 22.245546847367528,
|
| 756 |
+
"learning_rate": 1.8754043638935283e-05,
|
| 757 |
+
"loss": 1.2532,
|
| 758 |
+
"step": 530
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.24699907663896584,
|
| 762 |
+
"grad_norm": 4.177323522295811,
|
| 763 |
+
"learning_rate": 1.871480082998371e-05,
|
| 764 |
+
"loss": 1.2501,
|
| 765 |
+
"step": 535
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.24930747922437674,
|
| 769 |
+
"grad_norm": 3.9426463777773346,
|
| 770 |
+
"learning_rate": 1.867499194966106e-05,
|
| 771 |
+
"loss": 1.2683,
|
| 772 |
+
"step": 540
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.2516158818097876,
|
| 776 |
+
"grad_norm": 3.912690873331932,
|
| 777 |
+
"learning_rate": 1.8634619583759933e-05,
|
| 778 |
+
"loss": 1.2874,
|
| 779 |
+
"step": 545
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.25392428439519854,
|
| 783 |
+
"grad_norm": 3.972529239438344,
|
| 784 |
+
"learning_rate": 1.8593686354674223e-05,
|
| 785 |
+
"loss": 1.2698,
|
| 786 |
+
"step": 550
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.2562326869806094,
|
| 790 |
+
"grad_norm": 3.958572886167977,
|
| 791 |
+
"learning_rate": 1.8552194921228793e-05,
|
| 792 |
+
"loss": 1.2293,
|
| 793 |
+
"step": 555
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.2585410895660203,
|
| 797 |
+
"grad_norm": 3.7553829117034767,
|
| 798 |
+
"learning_rate": 1.851014797850676e-05,
|
| 799 |
+
"loss": 1.2818,
|
| 800 |
+
"step": 560
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2608494921514312,
|
| 804 |
+
"grad_norm": 4.352268879736511,
|
| 805 |
+
"learning_rate": 1.8467548257674453e-05,
|
| 806 |
+
"loss": 1.2552,
|
| 807 |
+
"step": 565
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2631578947368421,
|
| 811 |
+
"grad_norm": 5.014139215739045,
|
| 812 |
+
"learning_rate": 1.8424398525803983e-05,
|
| 813 |
+
"loss": 1.2228,
|
| 814 |
+
"step": 570
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.265466297322253,
|
| 818 |
+
"grad_norm": 4.192590762422093,
|
| 819 |
+
"learning_rate": 1.8380701585693526e-05,
|
| 820 |
+
"loss": 1.2526,
|
| 821 |
+
"step": 575
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2677746999076639,
|
| 825 |
+
"grad_norm": 4.209340122955672,
|
| 826 |
+
"learning_rate": 1.8336460275685267e-05,
|
| 827 |
+
"loss": 1.2681,
|
| 828 |
+
"step": 580
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.27008310249307477,
|
| 832 |
+
"grad_norm": 3.801129619164067,
|
| 833 |
+
"learning_rate": 1.8291677469481025e-05,
|
| 834 |
+
"loss": 1.2623,
|
| 835 |
+
"step": 585
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2723915050784857,
|
| 839 |
+
"grad_norm": 5.60448449703679,
|
| 840 |
+
"learning_rate": 1.8246356075955594e-05,
|
| 841 |
+
"loss": 1.2778,
|
| 842 |
+
"step": 590
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.27469990766389657,
|
| 846 |
+
"grad_norm": 3.8415685450636143,
|
| 847 |
+
"learning_rate": 1.820049903896782e-05,
|
| 848 |
+
"loss": 1.2546,
|
| 849 |
+
"step": 595
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.2770083102493075,
|
| 853 |
+
"grad_norm": 3.766423848242755,
|
| 854 |
+
"learning_rate": 1.8154109337169326e-05,
|
| 855 |
+
"loss": 1.2994,
|
| 856 |
+
"step": 600
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.2793167128347184,
|
| 860 |
+
"grad_norm": 3.8445299977202363,
|
| 861 |
+
"learning_rate": 1.8107189983811094e-05,
|
| 862 |
+
"loss": 1.2779,
|
| 863 |
+
"step": 605
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.28162511542012925,
|
| 867 |
+
"grad_norm": 4.20182793655244,
|
| 868 |
+
"learning_rate": 1.8059744026547713e-05,
|
| 869 |
+
"loss": 1.2794,
|
| 870 |
+
"step": 610
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2839335180055402,
|
| 874 |
+
"grad_norm": 3.6927184982852554,
|
| 875 |
+
"learning_rate": 1.8011774547239403e-05,
|
| 876 |
+
"loss": 1.2217,
|
| 877 |
+
"step": 615
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.28624192059095105,
|
| 881 |
+
"grad_norm": 3.906241578603264,
|
| 882 |
+
"learning_rate": 1.796328466175186e-05,
|
| 883 |
+
"loss": 1.3162,
|
| 884 |
+
"step": 620
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.288550323176362,
|
| 888 |
+
"grad_norm": 3.7221850266429675,
|
| 889 |
+
"learning_rate": 1.791427751975385e-05,
|
| 890 |
+
"loss": 1.2591,
|
| 891 |
+
"step": 625
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.29085872576177285,
|
| 895 |
+
"grad_norm": 4.11815775927983,
|
| 896 |
+
"learning_rate": 1.786475630451262e-05,
|
| 897 |
+
"loss": 1.2572,
|
| 898 |
+
"step": 630
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.2931671283471837,
|
| 902 |
+
"grad_norm": 3.8995508626898454,
|
| 903 |
+
"learning_rate": 1.781472423268713e-05,
|
| 904 |
+
"loss": 1.2604,
|
| 905 |
+
"step": 635
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.29547553093259465,
|
| 909 |
+
"grad_norm": 4.5219499712986035,
|
| 910 |
+
"learning_rate": 1.776418455411913e-05,
|
| 911 |
+
"loss": 1.298,
|
| 912 |
+
"step": 640
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.29778393351800553,
|
| 916 |
+
"grad_norm": 4.5899598168207785,
|
| 917 |
+
"learning_rate": 1.7713140551622032e-05,
|
| 918 |
+
"loss": 1.2664,
|
| 919 |
+
"step": 645
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.30009233610341646,
|
| 923 |
+
"grad_norm": 4.641570078800192,
|
| 924 |
+
"learning_rate": 1.7661595540767714e-05,
|
| 925 |
+
"loss": 1.2689,
|
| 926 |
+
"step": 650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.30240073868882733,
|
| 930 |
+
"grad_norm": 4.383087991217795,
|
| 931 |
+
"learning_rate": 1.7609552869671126e-05,
|
| 932 |
+
"loss": 1.2551,
|
| 933 |
+
"step": 655
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3047091412742382,
|
| 937 |
+
"grad_norm": 3.9687899547292576,
|
| 938 |
+
"learning_rate": 1.7557015918772822e-05,
|
| 939 |
+
"loss": 1.2379,
|
| 940 |
+
"step": 660
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.30701754385964913,
|
| 944 |
+
"grad_norm": 4.133840300932013,
|
| 945 |
+
"learning_rate": 1.750398810061939e-05,
|
| 946 |
+
"loss": 1.2779,
|
| 947 |
+
"step": 665
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.30932594644506,
|
| 951 |
+
"grad_norm": 3.84778329275165,
|
| 952 |
+
"learning_rate": 1.745047285964179e-05,
|
| 953 |
+
"loss": 1.2306,
|
| 954 |
+
"step": 670
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.31163434903047094,
|
| 958 |
+
"grad_norm": 4.054603771464119,
|
| 959 |
+
"learning_rate": 1.7396473671931597e-05,
|
| 960 |
+
"loss": 1.2089,
|
| 961 |
+
"step": 675
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.3139427516158818,
|
| 965 |
+
"grad_norm": 4.013882196193361,
|
| 966 |
+
"learning_rate": 1.7341994045015245e-05,
|
| 967 |
+
"loss": 1.2225,
|
| 968 |
+
"step": 680
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.3162511542012927,
|
| 972 |
+
"grad_norm": 4.076399340438248,
|
| 973 |
+
"learning_rate": 1.7287037517626174e-05,
|
| 974 |
+
"loss": 1.3166,
|
| 975 |
+
"step": 685
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.3185595567867036,
|
| 979 |
+
"grad_norm": 3.991144267549364,
|
| 980 |
+
"learning_rate": 1.7231607659474972e-05,
|
| 981 |
+
"loss": 1.2706,
|
| 982 |
+
"step": 690
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.3208679593721145,
|
| 986 |
+
"grad_norm": 3.592102167186549,
|
| 987 |
+
"learning_rate": 1.7175708071017503e-05,
|
| 988 |
+
"loss": 1.2066,
|
| 989 |
+
"step": 695
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.3231763619575254,
|
| 993 |
+
"grad_norm": 4.2490266329322655,
|
| 994 |
+
"learning_rate": 1.7119342383221055e-05,
|
| 995 |
+
"loss": 1.3011,
|
| 996 |
+
"step": 700
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.3254847645429363,
|
| 1000 |
+
"grad_norm": 3.7487591296204266,
|
| 1001 |
+
"learning_rate": 1.7062514257328474e-05,
|
| 1002 |
+
"loss": 1.2587,
|
| 1003 |
+
"step": 705
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.32779316712834716,
|
| 1007 |
+
"grad_norm": 3.6111287365523466,
|
| 1008 |
+
"learning_rate": 1.7005227384620336e-05,
|
| 1009 |
+
"loss": 1.2626,
|
| 1010 |
+
"step": 710
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3301015697137581,
|
| 1014 |
+
"grad_norm": 3.8624035554609892,
|
| 1015 |
+
"learning_rate": 1.6947485486175223e-05,
|
| 1016 |
+
"loss": 1.266,
|
| 1017 |
+
"step": 715
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.33240997229916897,
|
| 1021 |
+
"grad_norm": 4.191574332500623,
|
| 1022 |
+
"learning_rate": 1.688929231262797e-05,
|
| 1023 |
+
"loss": 1.2275,
|
| 1024 |
+
"step": 720
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3347183748845799,
|
| 1028 |
+
"grad_norm": 3.931766819485826,
|
| 1029 |
+
"learning_rate": 1.683065164392606e-05,
|
| 1030 |
+
"loss": 1.2525,
|
| 1031 |
+
"step": 725
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.33702677746999077,
|
| 1035 |
+
"grad_norm": 3.8224846577065685,
|
| 1036 |
+
"learning_rate": 1.6771567289084122e-05,
|
| 1037 |
+
"loss": 1.228,
|
| 1038 |
+
"step": 730
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.33933518005540164,
|
| 1042 |
+
"grad_norm": 3.7975499971303024,
|
| 1043 |
+
"learning_rate": 1.6712043085936473e-05,
|
| 1044 |
+
"loss": 1.2121,
|
| 1045 |
+
"step": 735
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.34164358264081257,
|
| 1049 |
+
"grad_norm": 3.7233983105114326,
|
| 1050 |
+
"learning_rate": 1.6652082900887858e-05,
|
| 1051 |
+
"loss": 1.2439,
|
| 1052 |
+
"step": 740
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.34395198522622344,
|
| 1056 |
+
"grad_norm": 4.0496534376278674,
|
| 1057 |
+
"learning_rate": 1.6591690628662305e-05,
|
| 1058 |
+
"loss": 1.3064,
|
| 1059 |
+
"step": 745
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3462603878116344,
|
| 1063 |
+
"grad_norm": 4.397682055950332,
|
| 1064 |
+
"learning_rate": 1.6530870192050134e-05,
|
| 1065 |
+
"loss": 1.2433,
|
| 1066 |
+
"step": 750
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.34856879039704525,
|
| 1070 |
+
"grad_norm": 3.999160650641557,
|
| 1071 |
+
"learning_rate": 1.6469625541653152e-05,
|
| 1072 |
+
"loss": 1.2117,
|
| 1073 |
+
"step": 755
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3508771929824561,
|
| 1077 |
+
"grad_norm": 4.475385002364299,
|
| 1078 |
+
"learning_rate": 1.6407960655628055e-05,
|
| 1079 |
+
"loss": 1.203,
|
| 1080 |
+
"step": 760
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.35318559556786705,
|
| 1084 |
+
"grad_norm": 3.5042875341184416,
|
| 1085 |
+
"learning_rate": 1.6345879539428e-05,
|
| 1086 |
+
"loss": 1.2567,
|
| 1087 |
+
"step": 765
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.3554939981532779,
|
| 1091 |
+
"grad_norm": 3.678612416780679,
|
| 1092 |
+
"learning_rate": 1.6283386225542467e-05,
|
| 1093 |
+
"loss": 1.2276,
|
| 1094 |
+
"step": 770
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.35780240073868885,
|
| 1098 |
+
"grad_norm": 5.063348081613382,
|
| 1099 |
+
"learning_rate": 1.622048477323529e-05,
|
| 1100 |
+
"loss": 1.2297,
|
| 1101 |
+
"step": 775
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.3601108033240997,
|
| 1105 |
+
"grad_norm": 4.04397764374825,
|
| 1106 |
+
"learning_rate": 1.6157179268281007e-05,
|
| 1107 |
+
"loss": 1.2498,
|
| 1108 |
+
"step": 780
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.3624192059095106,
|
| 1112 |
+
"grad_norm": 3.7786600086660553,
|
| 1113 |
+
"learning_rate": 1.6093473822699467e-05,
|
| 1114 |
+
"loss": 1.2156,
|
| 1115 |
+
"step": 785
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.36472760849492153,
|
| 1119 |
+
"grad_norm": 3.726670143436363,
|
| 1120 |
+
"learning_rate": 1.6029372574488732e-05,
|
| 1121 |
+
"loss": 1.248,
|
| 1122 |
+
"step": 790
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.3670360110803324,
|
| 1126 |
+
"grad_norm": 3.6023664901819115,
|
| 1127 |
+
"learning_rate": 1.5964879687356286e-05,
|
| 1128 |
+
"loss": 1.2762,
|
| 1129 |
+
"step": 795
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.36934441366574333,
|
| 1133 |
+
"grad_norm": 3.684618843127009,
|
| 1134 |
+
"learning_rate": 1.589999935044859e-05,
|
| 1135 |
+
"loss": 1.2269,
|
| 1136 |
+
"step": 800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.3716528162511542,
|
| 1140 |
+
"grad_norm": 3.6119834291134465,
|
| 1141 |
+
"learning_rate": 1.5834735778078968e-05,
|
| 1142 |
+
"loss": 1.2078,
|
| 1143 |
+
"step": 805
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.3739612188365651,
|
| 1147 |
+
"grad_norm": 3.66332363718426,
|
| 1148 |
+
"learning_rate": 1.5769093209453876e-05,
|
| 1149 |
+
"loss": 1.2713,
|
| 1150 |
+
"step": 810
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.376269621421976,
|
| 1154 |
+
"grad_norm": 4.137676249046753,
|
| 1155 |
+
"learning_rate": 1.5703075908397523e-05,
|
| 1156 |
+
"loss": 1.2816,
|
| 1157 |
+
"step": 815
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.3785780240073869,
|
| 1161 |
+
"grad_norm": 3.8481468093108475,
|
| 1162 |
+
"learning_rate": 1.563668816307494e-05,
|
| 1163 |
+
"loss": 1.2203,
|
| 1164 |
+
"step": 820
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3808864265927978,
|
| 1168 |
+
"grad_norm": 3.7158307301305156,
|
| 1169 |
+
"learning_rate": 1.556993428571342e-05,
|
| 1170 |
+
"loss": 1.2163,
|
| 1171 |
+
"step": 825
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3831948291782087,
|
| 1175 |
+
"grad_norm": 3.851222452502614,
|
| 1176 |
+
"learning_rate": 1.550281861232243e-05,
|
| 1177 |
+
"loss": 1.243,
|
| 1178 |
+
"step": 830
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.38550323176361956,
|
| 1182 |
+
"grad_norm": 3.6817891692377978,
|
| 1183 |
+
"learning_rate": 1.5435345502411956e-05,
|
| 1184 |
+
"loss": 1.2821,
|
| 1185 |
+
"step": 835
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3878116343490305,
|
| 1189 |
+
"grad_norm": 3.9683025462284998,
|
| 1190 |
+
"learning_rate": 1.536751933870934e-05,
|
| 1191 |
+
"loss": 1.2019,
|
| 1192 |
+
"step": 840
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.39012003693444136,
|
| 1196 |
+
"grad_norm": 3.94265762295689,
|
| 1197 |
+
"learning_rate": 1.5299344526874576e-05,
|
| 1198 |
+
"loss": 1.2774,
|
| 1199 |
+
"step": 845
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.39242843951985223,
|
| 1203 |
+
"grad_norm": 4.123641725136207,
|
| 1204 |
+
"learning_rate": 1.5230825495214184e-05,
|
| 1205 |
+
"loss": 1.2352,
|
| 1206 |
+
"step": 850
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.39473684210526316,
|
| 1210 |
+
"grad_norm": 3.9570109790957653,
|
| 1211 |
+
"learning_rate": 1.5161966694393516e-05,
|
| 1212 |
+
"loss": 1.215,
|
| 1213 |
+
"step": 855
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.39704524469067404,
|
| 1217 |
+
"grad_norm": 3.6427091867450714,
|
| 1218 |
+
"learning_rate": 1.5092772597147707e-05,
|
| 1219 |
+
"loss": 1.2202,
|
| 1220 |
+
"step": 860
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.39935364727608497,
|
| 1224 |
+
"grad_norm": 3.8425754107191796,
|
| 1225 |
+
"learning_rate": 1.5023247697991114e-05,
|
| 1226 |
+
"loss": 1.2432,
|
| 1227 |
+
"step": 865
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.40166204986149584,
|
| 1231 |
+
"grad_norm": 3.759319372367797,
|
| 1232 |
+
"learning_rate": 1.4953396512925398e-05,
|
| 1233 |
+
"loss": 1.1838,
|
| 1234 |
+
"step": 870
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.4039704524469067,
|
| 1238 |
+
"grad_norm": 3.872324982369786,
|
| 1239 |
+
"learning_rate": 1.4883223579146167e-05,
|
| 1240 |
+
"loss": 1.2331,
|
| 1241 |
+
"step": 875
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.40627885503231764,
|
| 1245 |
+
"grad_norm": 3.8616658245003435,
|
| 1246 |
+
"learning_rate": 1.4812733454748283e-05,
|
| 1247 |
+
"loss": 1.2277,
|
| 1248 |
+
"step": 880
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.4085872576177285,
|
| 1252 |
+
"grad_norm": 3.5624714154298163,
|
| 1253 |
+
"learning_rate": 1.4741930718429772e-05,
|
| 1254 |
+
"loss": 1.2051,
|
| 1255 |
+
"step": 885
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.41089566020313945,
|
| 1259 |
+
"grad_norm": 3.6961173549363924,
|
| 1260 |
+
"learning_rate": 1.4670819969194416e-05,
|
| 1261 |
+
"loss": 1.2309,
|
| 1262 |
+
"step": 890
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.4132040627885503,
|
| 1266 |
+
"grad_norm": 3.5654510220296847,
|
| 1267 |
+
"learning_rate": 1.4599405826053039e-05,
|
| 1268 |
+
"loss": 1.1884,
|
| 1269 |
+
"step": 895
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.4155124653739612,
|
| 1273 |
+
"grad_norm": 4.205884899208378,
|
| 1274 |
+
"learning_rate": 1.4527692927723465e-05,
|
| 1275 |
+
"loss": 1.2223,
|
| 1276 |
+
"step": 900
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.4178208679593721,
|
| 1280 |
+
"grad_norm": 3.9431786244545997,
|
| 1281 |
+
"learning_rate": 1.4455685932329204e-05,
|
| 1282 |
+
"loss": 1.2389,
|
| 1283 |
+
"step": 905
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.420129270544783,
|
| 1287 |
+
"grad_norm": 3.579703652121505,
|
| 1288 |
+
"learning_rate": 1.4383389517096899e-05,
|
| 1289 |
+
"loss": 1.2429,
|
| 1290 |
+
"step": 910
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.4224376731301939,
|
| 1294 |
+
"grad_norm": 3.7807582830713105,
|
| 1295 |
+
"learning_rate": 1.4310808378052506e-05,
|
| 1296 |
+
"loss": 1.1874,
|
| 1297 |
+
"step": 915
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.4247460757156048,
|
| 1301 |
+
"grad_norm": 3.9020463886513914,
|
| 1302 |
+
"learning_rate": 1.4237947229716262e-05,
|
| 1303 |
+
"loss": 1.2587,
|
| 1304 |
+
"step": 920
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.42705447830101567,
|
| 1308 |
+
"grad_norm": 3.7663448915088633,
|
| 1309 |
+
"learning_rate": 1.4164810804796464e-05,
|
| 1310 |
+
"loss": 1.184,
|
| 1311 |
+
"step": 925
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.4293628808864266,
|
| 1315 |
+
"grad_norm": 3.7907471270783937,
|
| 1316 |
+
"learning_rate": 1.409140385388203e-05,
|
| 1317 |
+
"loss": 1.2445,
|
| 1318 |
+
"step": 930
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.4316712834718375,
|
| 1322 |
+
"grad_norm": 3.791543245723202,
|
| 1323 |
+
"learning_rate": 1.4017731145133955e-05,
|
| 1324 |
+
"loss": 1.2527,
|
| 1325 |
+
"step": 935
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.4339796860572484,
|
| 1329 |
+
"grad_norm": 3.8566751713668666,
|
| 1330 |
+
"learning_rate": 1.3943797463975575e-05,
|
| 1331 |
+
"loss": 1.2048,
|
| 1332 |
+
"step": 940
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.4362880886426593,
|
| 1336 |
+
"grad_norm": 3.943257567360323,
|
| 1337 |
+
"learning_rate": 1.3869607612781733e-05,
|
| 1338 |
+
"loss": 1.2773,
|
| 1339 |
+
"step": 945
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.43859649122807015,
|
| 1343 |
+
"grad_norm": 3.53206021655625,
|
| 1344 |
+
"learning_rate": 1.3795166410566834e-05,
|
| 1345 |
+
"loss": 1.2066,
|
| 1346 |
+
"step": 950
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.4409048938134811,
|
| 1350 |
+
"grad_norm": 3.8322607840339504,
|
| 1351 |
+
"learning_rate": 1.372047869267184e-05,
|
| 1352 |
+
"loss": 1.2104,
|
| 1353 |
+
"step": 955
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.44321329639889195,
|
| 1357 |
+
"grad_norm": 4.982802180271467,
|
| 1358 |
+
"learning_rate": 1.364554931045018e-05,
|
| 1359 |
+
"loss": 1.2782,
|
| 1360 |
+
"step": 960
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.4455216989843029,
|
| 1364 |
+
"grad_norm": 4.121927772157904,
|
| 1365 |
+
"learning_rate": 1.3570383130952627e-05,
|
| 1366 |
+
"loss": 1.2221,
|
| 1367 |
+
"step": 965
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.44783010156971376,
|
| 1371 |
+
"grad_norm": 3.5401426054616674,
|
| 1372 |
+
"learning_rate": 1.349498503661116e-05,
|
| 1373 |
+
"loss": 1.249,
|
| 1374 |
+
"step": 970
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.45013850415512463,
|
| 1378 |
+
"grad_norm": 3.8347876039826647,
|
| 1379 |
+
"learning_rate": 1.3419359924921833e-05,
|
| 1380 |
+
"loss": 1.2736,
|
| 1381 |
+
"step": 975
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.45244690674053556,
|
| 1385 |
+
"grad_norm": 4.86416192250325,
|
| 1386 |
+
"learning_rate": 1.3343512708126642e-05,
|
| 1387 |
+
"loss": 1.2032,
|
| 1388 |
+
"step": 980
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.45475530932594643,
|
| 1392 |
+
"grad_norm": 3.8508803970513004,
|
| 1393 |
+
"learning_rate": 1.326744831289447e-05,
|
| 1394 |
+
"loss": 1.2465,
|
| 1395 |
+
"step": 985
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.45706371191135736,
|
| 1399 |
+
"grad_norm": 3.276661833625774,
|
| 1400 |
+
"learning_rate": 1.3191171680001048e-05,
|
| 1401 |
+
"loss": 1.1905,
|
| 1402 |
+
"step": 990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.45937211449676824,
|
| 1406 |
+
"grad_norm": 3.6488550777243933,
|
| 1407 |
+
"learning_rate": 1.3114687764008048e-05,
|
| 1408 |
+
"loss": 1.1991,
|
| 1409 |
+
"step": 995
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.4616805170821791,
|
| 1413 |
+
"grad_norm": 3.9637997706000223,
|
| 1414 |
+
"learning_rate": 1.3038001532941249e-05,
|
| 1415 |
+
"loss": 1.1994,
|
| 1416 |
+
"step": 1000
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.46398891966759004,
|
| 1420 |
+
"grad_norm": 3.7798295608326447,
|
| 1421 |
+
"learning_rate": 1.2961117967967844e-05,
|
| 1422 |
+
"loss": 1.2327,
|
| 1423 |
+
"step": 1005
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.4662973222530009,
|
| 1427 |
+
"grad_norm": 3.742363753899004,
|
| 1428 |
+
"learning_rate": 1.2884042063072881e-05,
|
| 1429 |
+
"loss": 1.2415,
|
| 1430 |
+
"step": 1010
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.46860572483841184,
|
| 1434 |
+
"grad_norm": 4.00995610689072,
|
| 1435 |
+
"learning_rate": 1.280677882473488e-05,
|
| 1436 |
+
"loss": 1.2449,
|
| 1437 |
+
"step": 1015
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.4709141274238227,
|
| 1441 |
+
"grad_norm": 3.7802768150285284,
|
| 1442 |
+
"learning_rate": 1.272933327160063e-05,
|
| 1443 |
+
"loss": 1.2055,
|
| 1444 |
+
"step": 1020
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.4732225300092336,
|
| 1448 |
+
"grad_norm": 3.979719082398227,
|
| 1449 |
+
"learning_rate": 1.2651710434159223e-05,
|
| 1450 |
+
"loss": 1.1452,
|
| 1451 |
+
"step": 1025
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.4755309325946445,
|
| 1455 |
+
"grad_norm": 3.7987734509998012,
|
| 1456 |
+
"learning_rate": 1.2573915354415274e-05,
|
| 1457 |
+
"loss": 1.2266,
|
| 1458 |
+
"step": 1030
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.4778393351800554,
|
| 1462 |
+
"grad_norm": 3.4449265105850344,
|
| 1463 |
+
"learning_rate": 1.2495953085561426e-05,
|
| 1464 |
+
"loss": 1.1678,
|
| 1465 |
+
"step": 1035
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.4801477377654663,
|
| 1469 |
+
"grad_norm": 4.703831538180476,
|
| 1470 |
+
"learning_rate": 1.241782869165012e-05,
|
| 1471 |
+
"loss": 1.1893,
|
| 1472 |
+
"step": 1040
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.4824561403508772,
|
| 1476 |
+
"grad_norm": 3.56138065098868,
|
| 1477 |
+
"learning_rate": 1.2339547247264658e-05,
|
| 1478 |
+
"loss": 1.2285,
|
| 1479 |
+
"step": 1045
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.48476454293628807,
|
| 1483 |
+
"grad_norm": 3.8664090630676147,
|
| 1484 |
+
"learning_rate": 1.2261113837189587e-05,
|
| 1485 |
+
"loss": 1.1995,
|
| 1486 |
+
"step": 1050
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.487072945521699,
|
| 1490 |
+
"grad_norm": 3.6587622685467553,
|
| 1491 |
+
"learning_rate": 1.2182533556080402e-05,
|
| 1492 |
+
"loss": 1.2456,
|
| 1493 |
+
"step": 1055
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.48938134810710987,
|
| 1497 |
+
"grad_norm": 3.4219623018934615,
|
| 1498 |
+
"learning_rate": 1.2103811508132642e-05,
|
| 1499 |
+
"loss": 1.1904,
|
| 1500 |
+
"step": 1060
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.4916897506925208,
|
| 1504 |
+
"grad_norm": 3.91141223990254,
|
| 1505 |
+
"learning_rate": 1.2024952806750321e-05,
|
| 1506 |
+
"loss": 1.1811,
|
| 1507 |
+
"step": 1065
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.4939981532779317,
|
| 1511 |
+
"grad_norm": 3.707066130468398,
|
| 1512 |
+
"learning_rate": 1.1945962574213814e-05,
|
| 1513 |
+
"loss": 1.212,
|
| 1514 |
+
"step": 1070
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.49630655586334255,
|
| 1518 |
+
"grad_norm": 3.5782501836947653,
|
| 1519 |
+
"learning_rate": 1.1866845941347118e-05,
|
| 1520 |
+
"loss": 1.2255,
|
| 1521 |
+
"step": 1075
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.4986149584487535,
|
| 1525 |
+
"grad_norm": 4.303350644777213,
|
| 1526 |
+
"learning_rate": 1.1787608047184583e-05,
|
| 1527 |
+
"loss": 1.1376,
|
| 1528 |
+
"step": 1080
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5009233610341643,
|
| 1532 |
+
"grad_norm": 3.419543860379626,
|
| 1533 |
+
"learning_rate": 1.1708254038637115e-05,
|
| 1534 |
+
"loss": 1.1872,
|
| 1535 |
+
"step": 1085
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5032317636195752,
|
| 1539 |
+
"grad_norm": 3.586294780528409,
|
| 1540 |
+
"learning_rate": 1.1628789070157836e-05,
|
| 1541 |
+
"loss": 1.2114,
|
| 1542 |
+
"step": 1090
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5055401662049861,
|
| 1546 |
+
"grad_norm": 3.6647616517214496,
|
| 1547 |
+
"learning_rate": 1.1549218303407305e-05,
|
| 1548 |
+
"loss": 1.2088,
|
| 1549 |
+
"step": 1095
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5078485687903971,
|
| 1553 |
+
"grad_norm": 3.6209405687157794,
|
| 1554 |
+
"learning_rate": 1.1469546906918219e-05,
|
| 1555 |
+
"loss": 1.1535,
|
| 1556 |
+
"step": 1100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.510156971375808,
|
| 1560 |
+
"grad_norm": 3.4760951984933777,
|
| 1561 |
+
"learning_rate": 1.1389780055759689e-05,
|
| 1562 |
+
"loss": 1.1692,
|
| 1563 |
+
"step": 1105
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5124653739612188,
|
| 1567 |
+
"grad_norm": 3.523587148397925,
|
| 1568 |
+
"learning_rate": 1.1309922931201114e-05,
|
| 1569 |
+
"loss": 1.1795,
|
| 1570 |
+
"step": 1110
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5147737765466297,
|
| 1574 |
+
"grad_norm": 3.399747435026194,
|
| 1575 |
+
"learning_rate": 1.1229980720375609e-05,
|
| 1576 |
+
"loss": 1.1913,
|
| 1577 |
+
"step": 1115
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5170821791320406,
|
| 1581 |
+
"grad_norm": 3.802970464768176,
|
| 1582 |
+
"learning_rate": 1.114995861594308e-05,
|
| 1583 |
+
"loss": 1.1692,
|
| 1584 |
+
"step": 1120
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.5193905817174516,
|
| 1588 |
+
"grad_norm": 3.571347595436078,
|
| 1589 |
+
"learning_rate": 1.1069861815752944e-05,
|
| 1590 |
+
"loss": 1.1575,
|
| 1591 |
+
"step": 1125
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.5216989843028624,
|
| 1595 |
+
"grad_norm": 3.702241350827994,
|
| 1596 |
+
"learning_rate": 1.0989695522506486e-05,
|
| 1597 |
+
"loss": 1.1776,
|
| 1598 |
+
"step": 1130
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5240073868882733,
|
| 1602 |
+
"grad_norm": 4.396145181294285,
|
| 1603 |
+
"learning_rate": 1.0909464943418926e-05,
|
| 1604 |
+
"loss": 1.2055,
|
| 1605 |
+
"step": 1135
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5263157894736842,
|
| 1609 |
+
"grad_norm": 3.402649511273165,
|
| 1610 |
+
"learning_rate": 1.0829175289881188e-05,
|
| 1611 |
+
"loss": 1.2024,
|
| 1612 |
+
"step": 1140
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.528624192059095,
|
| 1616 |
+
"grad_norm": 3.321901777095843,
|
| 1617 |
+
"learning_rate": 1.074883177712138e-05,
|
| 1618 |
+
"loss": 1.1317,
|
| 1619 |
+
"step": 1145
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.530932594644506,
|
| 1623 |
+
"grad_norm": 4.575011114858196,
|
| 1624 |
+
"learning_rate": 1.0668439623866043e-05,
|
| 1625 |
+
"loss": 1.1516,
|
| 1626 |
+
"step": 1150
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5332409972299169,
|
| 1630 |
+
"grad_norm": 3.428811319179132,
|
| 1631 |
+
"learning_rate": 1.0588004052001177e-05,
|
| 1632 |
+
"loss": 1.1326,
|
| 1633 |
+
"step": 1155
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5355493998153278,
|
| 1637 |
+
"grad_norm": 3.758823500740248,
|
| 1638 |
+
"learning_rate": 1.0507530286233042e-05,
|
| 1639 |
+
"loss": 1.1523,
|
| 1640 |
+
"step": 1160
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.5378578024007387,
|
| 1644 |
+
"grad_norm": 3.828420445656179,
|
| 1645 |
+
"learning_rate": 1.0427023553748792e-05,
|
| 1646 |
+
"loss": 1.215,
|
| 1647 |
+
"step": 1165
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5401662049861495,
|
| 1651 |
+
"grad_norm": 3.872474623427253,
|
| 1652 |
+
"learning_rate": 1.0346489083876928e-05,
|
| 1653 |
+
"loss": 1.1798,
|
| 1654 |
+
"step": 1170
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.5424746075715605,
|
| 1658 |
+
"grad_norm": 4.343223419966708,
|
| 1659 |
+
"learning_rate": 1.0265932107747656e-05,
|
| 1660 |
+
"loss": 1.1964,
|
| 1661 |
+
"step": 1175
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.5447830101569714,
|
| 1665 |
+
"grad_norm": 3.4458152638291533,
|
| 1666 |
+
"learning_rate": 1.0185357857953064e-05,
|
| 1667 |
+
"loss": 1.188,
|
| 1668 |
+
"step": 1180
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.5470914127423823,
|
| 1672 |
+
"grad_norm": 3.3343026801443765,
|
| 1673 |
+
"learning_rate": 1.0104771568207266e-05,
|
| 1674 |
+
"loss": 1.1524,
|
| 1675 |
+
"step": 1185
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.5493998153277931,
|
| 1679 |
+
"grad_norm": 3.8325280372919774,
|
| 1680 |
+
"learning_rate": 1.0024178473006418e-05,
|
| 1681 |
+
"loss": 1.1445,
|
| 1682 |
+
"step": 1190
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.551708217913204,
|
| 1686 |
+
"grad_norm": 3.913934401934443,
|
| 1687 |
+
"learning_rate": 9.943583807288746e-06,
|
| 1688 |
+
"loss": 1.1497,
|
| 1689 |
+
"step": 1195
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.554016620498615,
|
| 1693 |
+
"grad_norm": 3.8771337742661585,
|
| 1694 |
+
"learning_rate": 9.862992806094473e-06,
|
| 1695 |
+
"loss": 1.1584,
|
| 1696 |
+
"step": 1200
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5563250230840259,
|
| 1700 |
+
"grad_norm": 3.385706053842486,
|
| 1701 |
+
"learning_rate": 9.782410704225793e-06,
|
| 1702 |
+
"loss": 1.133,
|
| 1703 |
+
"step": 1205
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5586334256694367,
|
| 1707 |
+
"grad_norm": 3.228558572718497,
|
| 1708 |
+
"learning_rate": 9.701842735906855e-06,
|
| 1709 |
+
"loss": 1.1714,
|
| 1710 |
+
"step": 1210
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5609418282548476,
|
| 1714 |
+
"grad_norm": 3.376489834368575,
|
| 1715 |
+
"learning_rate": 9.621294134443747e-06,
|
| 1716 |
+
"loss": 1.1782,
|
| 1717 |
+
"step": 1215
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5632502308402585,
|
| 1721 |
+
"grad_norm": 4.101023970778267,
|
| 1722 |
+
"learning_rate": 9.54077013188459e-06,
|
| 1723 |
+
"loss": 1.1679,
|
| 1724 |
+
"step": 1220
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5655586334256695,
|
| 1728 |
+
"grad_norm": 3.459693677788322,
|
| 1729 |
+
"learning_rate": 9.460275958679674e-06,
|
| 1730 |
+
"loss": 1.2272,
|
| 1731 |
+
"step": 1225
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5678670360110804,
|
| 1735 |
+
"grad_norm": 3.5741244509053556,
|
| 1736 |
+
"learning_rate": 9.379816843341715e-06,
|
| 1737 |
+
"loss": 1.1679,
|
| 1738 |
+
"step": 1230
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5701754385964912,
|
| 1742 |
+
"grad_norm": 14.959841662019736,
|
| 1743 |
+
"learning_rate": 9.299398012106246e-06,
|
| 1744 |
+
"loss": 1.1557,
|
| 1745 |
+
"step": 1235
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5724838411819021,
|
| 1749 |
+
"grad_norm": 3.479142794568544,
|
| 1750 |
+
"learning_rate": 9.219024688592136e-06,
|
| 1751 |
+
"loss": 1.191,
|
| 1752 |
+
"step": 1240
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.574792243767313,
|
| 1756 |
+
"grad_norm": 3.4791994405128195,
|
| 1757 |
+
"learning_rate": 9.138702093462286e-06,
|
| 1758 |
+
"loss": 1.1632,
|
| 1759 |
+
"step": 1245
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.577100646352724,
|
| 1763 |
+
"grad_norm": 3.378297795269278,
|
| 1764 |
+
"learning_rate": 9.058435444084543e-06,
|
| 1765 |
+
"loss": 1.2058,
|
| 1766 |
+
"step": 1250
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5794090489381348,
|
| 1770 |
+
"grad_norm": 3.3312286796444948,
|
| 1771 |
+
"learning_rate": 8.978229954192775e-06,
|
| 1772 |
+
"loss": 1.2072,
|
| 1773 |
+
"step": 1255
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5817174515235457,
|
| 1777 |
+
"grad_norm": 3.2936946867277497,
|
| 1778 |
+
"learning_rate": 8.898090833548226e-06,
|
| 1779 |
+
"loss": 1.1479,
|
| 1780 |
+
"step": 1260
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5840258541089566,
|
| 1784 |
+
"grad_norm": 3.5657195698986306,
|
| 1785 |
+
"learning_rate": 8.818023287601117e-06,
|
| 1786 |
+
"loss": 1.1579,
|
| 1787 |
+
"step": 1265
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5863342566943675,
|
| 1791 |
+
"grad_norm": 3.85534125869907,
|
| 1792 |
+
"learning_rate": 8.738032517152523e-06,
|
| 1793 |
+
"loss": 1.1748,
|
| 1794 |
+
"step": 1270
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5886426592797784,
|
| 1798 |
+
"grad_norm": 3.3807308381583585,
|
| 1799 |
+
"learning_rate": 8.658123718016548e-06,
|
| 1800 |
+
"loss": 1.1365,
|
| 1801 |
+
"step": 1275
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5909510618651893,
|
| 1805 |
+
"grad_norm": 3.75547737356039,
|
| 1806 |
+
"learning_rate": 8.578302080682844e-06,
|
| 1807 |
+
"loss": 1.1657,
|
| 1808 |
+
"step": 1280
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.5932594644506002,
|
| 1812 |
+
"grad_norm": 3.334058557259955,
|
| 1813 |
+
"learning_rate": 8.498572789979446e-06,
|
| 1814 |
+
"loss": 1.1653,
|
| 1815 |
+
"step": 1285
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5955678670360111,
|
| 1819 |
+
"grad_norm": 3.596795067568704,
|
| 1820 |
+
"learning_rate": 8.418941024735997e-06,
|
| 1821 |
+
"loss": 1.1909,
|
| 1822 |
+
"step": 1290
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5978762696214219,
|
| 1826 |
+
"grad_norm": 3.754106205642103,
|
| 1827 |
+
"learning_rate": 8.33941195744737e-06,
|
| 1828 |
+
"loss": 1.1595,
|
| 1829 |
+
"step": 1295
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.6001846722068329,
|
| 1833 |
+
"grad_norm": 3.3575559431036988,
|
| 1834 |
+
"learning_rate": 8.259990753937662e-06,
|
| 1835 |
+
"loss": 1.1378,
|
| 1836 |
+
"step": 1300
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6024930747922438,
|
| 1840 |
+
"grad_norm": 4.011372021010383,
|
| 1841 |
+
"learning_rate": 8.18068257302466e-06,
|
| 1842 |
+
"loss": 1.1832,
|
| 1843 |
+
"step": 1305
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.6048014773776547,
|
| 1847 |
+
"grad_norm": 3.404379906541828,
|
| 1848 |
+
"learning_rate": 8.101492566184757e-06,
|
| 1849 |
+
"loss": 1.1592,
|
| 1850 |
+
"step": 1310
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6071098799630655,
|
| 1854 |
+
"grad_norm": 3.498132121516362,
|
| 1855 |
+
"learning_rate": 8.022425877218321e-06,
|
| 1856 |
+
"loss": 1.1591,
|
| 1857 |
+
"step": 1315
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6094182825484764,
|
| 1861 |
+
"grad_norm": 3.523586580045349,
|
| 1862 |
+
"learning_rate": 7.943487641915595e-06,
|
| 1863 |
+
"loss": 1.1525,
|
| 1864 |
+
"step": 1320
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6117266851338874,
|
| 1868 |
+
"grad_norm": 3.6229726894839858,
|
| 1869 |
+
"learning_rate": 7.864682987723082e-06,
|
| 1870 |
+
"loss": 1.1618,
|
| 1871 |
+
"step": 1325
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6140350877192983,
|
| 1875 |
+
"grad_norm": 3.696989469787097,
|
| 1876 |
+
"learning_rate": 7.78601703341051e-06,
|
| 1877 |
+
"loss": 1.1824,
|
| 1878 |
+
"step": 1330
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.6163434903047091,
|
| 1882 |
+
"grad_norm": 3.567967173775001,
|
| 1883 |
+
"learning_rate": 7.70749488873833e-06,
|
| 1884 |
+
"loss": 1.1792,
|
| 1885 |
+
"step": 1335
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.61865189289012,
|
| 1889 |
+
"grad_norm": 3.399928397766497,
|
| 1890 |
+
"learning_rate": 7.629121654125808e-06,
|
| 1891 |
+
"loss": 1.1438,
|
| 1892 |
+
"step": 1340
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6209602954755309,
|
| 1896 |
+
"grad_norm": 3.6344006441397414,
|
| 1897 |
+
"learning_rate": 7.550902420319742e-06,
|
| 1898 |
+
"loss": 1.1591,
|
| 1899 |
+
"step": 1345
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6232686980609419,
|
| 1903 |
+
"grad_norm": 3.538106316840523,
|
| 1904 |
+
"learning_rate": 7.472842268063776e-06,
|
| 1905 |
+
"loss": 1.1311,
|
| 1906 |
+
"step": 1350
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6255771006463527,
|
| 1910 |
+
"grad_norm": 3.661558906665894,
|
| 1911 |
+
"learning_rate": 7.394946267768381e-06,
|
| 1912 |
+
"loss": 1.1621,
|
| 1913 |
+
"step": 1355
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6278855032317636,
|
| 1917 |
+
"grad_norm": 3.6197107279149954,
|
| 1918 |
+
"learning_rate": 7.317219479181517e-06,
|
| 1919 |
+
"loss": 1.1028,
|
| 1920 |
+
"step": 1360
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6301939058171745,
|
| 1924 |
+
"grad_norm": 3.4094252840241355,
|
| 1925 |
+
"learning_rate": 7.23966695105996e-06,
|
| 1926 |
+
"loss": 1.119,
|
| 1927 |
+
"step": 1365
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6325023084025854,
|
| 1931 |
+
"grad_norm": 3.4085855538144467,
|
| 1932 |
+
"learning_rate": 7.162293720841378e-06,
|
| 1933 |
+
"loss": 1.1438,
|
| 1934 |
+
"step": 1370
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6348107109879964,
|
| 1938 |
+
"grad_norm": 4.073406312500022,
|
| 1939 |
+
"learning_rate": 7.085104814317101e-06,
|
| 1940 |
+
"loss": 1.1729,
|
| 1941 |
+
"step": 1375
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6371191135734072,
|
| 1945 |
+
"grad_norm": 3.572178264074241,
|
| 1946 |
+
"learning_rate": 7.008105245305699e-06,
|
| 1947 |
+
"loss": 1.1661,
|
| 1948 |
+
"step": 1380
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.6394275161588181,
|
| 1952 |
+
"grad_norm": 3.81528951625221,
|
| 1953 |
+
"learning_rate": 6.931300015327274e-06,
|
| 1954 |
+
"loss": 1.1571,
|
| 1955 |
+
"step": 1385
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.641735918744229,
|
| 1959 |
+
"grad_norm": 3.2846636335941763,
|
| 1960 |
+
"learning_rate": 6.854694113278614e-06,
|
| 1961 |
+
"loss": 1.154,
|
| 1962 |
+
"step": 1390
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6440443213296398,
|
| 1966 |
+
"grad_norm": 3.2544013227776007,
|
| 1967 |
+
"learning_rate": 6.7782925151091224e-06,
|
| 1968 |
+
"loss": 1.0823,
|
| 1969 |
+
"step": 1395
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6463527239150508,
|
| 1973 |
+
"grad_norm": 3.482450898904014,
|
| 1974 |
+
"learning_rate": 6.702100183497613e-06,
|
| 1975 |
+
"loss": 1.1803,
|
| 1976 |
+
"step": 1400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6486611265004617,
|
| 1980 |
+
"grad_norm": 3.412256349030684,
|
| 1981 |
+
"learning_rate": 6.62612206752995e-06,
|
| 1982 |
+
"loss": 1.1643,
|
| 1983 |
+
"step": 1405
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.6509695290858726,
|
| 1987 |
+
"grad_norm": 3.75196899322532,
|
| 1988 |
+
"learning_rate": 6.550363102377588e-06,
|
| 1989 |
+
"loss": 1.1117,
|
| 1990 |
+
"step": 1410
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6532779316712835,
|
| 1994 |
+
"grad_norm": 3.3485189294016258,
|
| 1995 |
+
"learning_rate": 6.474828208976998e-06,
|
| 1996 |
+
"loss": 1.1466,
|
| 1997 |
+
"step": 1415
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6555863342566943,
|
| 2001 |
+
"grad_norm": 3.4421443761863104,
|
| 2002 |
+
"learning_rate": 6.3995222937100455e-06,
|
| 2003 |
+
"loss": 1.1468,
|
| 2004 |
+
"step": 1420
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6578947368421053,
|
| 2008 |
+
"grad_norm": 3.4653107797221683,
|
| 2009 |
+
"learning_rate": 6.324450248085265e-06,
|
| 2010 |
+
"loss": 1.1418,
|
| 2011 |
+
"step": 1425
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6602031394275162,
|
| 2015 |
+
"grad_norm": 3.450235228111911,
|
| 2016 |
+
"learning_rate": 6.249616948420161e-06,
|
| 2017 |
+
"loss": 1.1393,
|
| 2018 |
+
"step": 1430
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6625115420129271,
|
| 2022 |
+
"grad_norm": 3.648594332616919,
|
| 2023 |
+
"learning_rate": 6.175027255524446e-06,
|
| 2024 |
+
"loss": 1.1263,
|
| 2025 |
+
"step": 1435
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6648199445983379,
|
| 2029 |
+
"grad_norm": 3.50804118935427,
|
| 2030 |
+
"learning_rate": 6.100686014384315e-06,
|
| 2031 |
+
"loss": 1.1497,
|
| 2032 |
+
"step": 1440
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6671283471837488,
|
| 2036 |
+
"grad_norm": 3.407303145023877,
|
| 2037 |
+
"learning_rate": 6.026598053847743e-06,
|
| 2038 |
+
"loss": 1.1217,
|
| 2039 |
+
"step": 1445
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6694367497691598,
|
| 2043 |
+
"grad_norm": 3.6049741156075426,
|
| 2044 |
+
"learning_rate": 5.952768186310813e-06,
|
| 2045 |
+
"loss": 1.2134,
|
| 2046 |
+
"step": 1450
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6717451523545707,
|
| 2050 |
+
"grad_norm": 3.347553717603198,
|
| 2051 |
+
"learning_rate": 5.879201207405136e-06,
|
| 2052 |
+
"loss": 1.1189,
|
| 2053 |
+
"step": 1455
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6740535549399815,
|
| 2057 |
+
"grad_norm": 3.7624263901785087,
|
| 2058 |
+
"learning_rate": 5.805901895686344e-06,
|
| 2059 |
+
"loss": 1.1217,
|
| 2060 |
+
"step": 1460
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6763619575253924,
|
| 2064 |
+
"grad_norm": 3.6359056480115193,
|
| 2065 |
+
"learning_rate": 5.732875012323712e-06,
|
| 2066 |
+
"loss": 1.1275,
|
| 2067 |
+
"step": 1465
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6786703601108033,
|
| 2071 |
+
"grad_norm": 3.5660085050284946,
|
| 2072 |
+
"learning_rate": 5.660125300790873e-06,
|
| 2073 |
+
"loss": 1.153,
|
| 2074 |
+
"step": 1470
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6809787626962143,
|
| 2078 |
+
"grad_norm": 3.4188915438262946,
|
| 2079 |
+
"learning_rate": 5.58765748655772e-06,
|
| 2080 |
+
"loss": 1.126,
|
| 2081 |
+
"step": 1475
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6832871652816251,
|
| 2085 |
+
"grad_norm": 3.7409360766713995,
|
| 2086 |
+
"learning_rate": 5.5154762767834605e-06,
|
| 2087 |
+
"loss": 1.1312,
|
| 2088 |
+
"step": 1480
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.685595567867036,
|
| 2092 |
+
"grad_norm": 3.5176838276710787,
|
| 2093 |
+
"learning_rate": 5.443586360010859e-06,
|
| 2094 |
+
"loss": 1.118,
|
| 2095 |
+
"step": 1485
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6879039704524469,
|
| 2099 |
+
"grad_norm": 3.940112737940071,
|
| 2100 |
+
"learning_rate": 5.3719924058616975e-06,
|
| 2101 |
+
"loss": 1.1084,
|
| 2102 |
+
"step": 1490
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6902123730378578,
|
| 2106 |
+
"grad_norm": 3.5725656073039516,
|
| 2107 |
+
"learning_rate": 5.30069906473345e-06,
|
| 2108 |
+
"loss": 1.1462,
|
| 2109 |
+
"step": 1495
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.6925207756232687,
|
| 2113 |
+
"grad_norm": 3.585328985764251,
|
| 2114 |
+
"learning_rate": 5.2297109674972166e-06,
|
| 2115 |
+
"loss": 1.1275,
|
| 2116 |
+
"step": 1500
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6948291782086796,
|
| 2120 |
+
"grad_norm": 3.7150630899084276,
|
| 2121 |
+
"learning_rate": 5.159032725196946e-06,
|
| 2122 |
+
"loss": 1.1573,
|
| 2123 |
+
"step": 1505
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6971375807940905,
|
| 2127 |
+
"grad_norm": 3.4991531847637893,
|
| 2128 |
+
"learning_rate": 5.088668928749891e-06,
|
| 2129 |
+
"loss": 1.1339,
|
| 2130 |
+
"step": 1510
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6994459833795014,
|
| 2134 |
+
"grad_norm": 3.337315702796277,
|
| 2135 |
+
"learning_rate": 5.0186241486484245e-06,
|
| 2136 |
+
"loss": 1.1121,
|
| 2137 |
+
"step": 1515
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7017543859649122,
|
| 2141 |
+
"grad_norm": 3.166495977462906,
|
| 2142 |
+
"learning_rate": 4.948902934663158e-06,
|
| 2143 |
+
"loss": 1.1207,
|
| 2144 |
+
"step": 1520
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.7040627885503232,
|
| 2148 |
+
"grad_norm": 3.269883473204096,
|
| 2149 |
+
"learning_rate": 4.879509815547413e-06,
|
| 2150 |
+
"loss": 1.1067,
|
| 2151 |
+
"step": 1525
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7063711911357341,
|
| 2155 |
+
"grad_norm": 3.2549683491943138,
|
| 2156 |
+
"learning_rate": 4.810449298743051e-06,
|
| 2157 |
+
"loss": 1.0858,
|
| 2158 |
+
"step": 1530
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.708679593721145,
|
| 2162 |
+
"grad_norm": 3.673192940396545,
|
| 2163 |
+
"learning_rate": 4.741725870087693e-06,
|
| 2164 |
+
"loss": 1.1674,
|
| 2165 |
+
"step": 1535
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7109879963065558,
|
| 2169 |
+
"grad_norm": 3.295243146355197,
|
| 2170 |
+
"learning_rate": 4.673343993523347e-06,
|
| 2171 |
+
"loss": 1.1087,
|
| 2172 |
+
"step": 1540
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7132963988919667,
|
| 2176 |
+
"grad_norm": 3.4162872942710867,
|
| 2177 |
+
"learning_rate": 4.605308110806436e-06,
|
| 2178 |
+
"loss": 1.1224,
|
| 2179 |
+
"step": 1545
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7156048014773777,
|
| 2183 |
+
"grad_norm": 3.3989883160652865,
|
| 2184 |
+
"learning_rate": 4.537622641219309e-06,
|
| 2185 |
+
"loss": 1.1307,
|
| 2186 |
+
"step": 1550
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7179132040627886,
|
| 2190 |
+
"grad_norm": 3.2956559559454663,
|
| 2191 |
+
"learning_rate": 4.47029198128316e-06,
|
| 2192 |
+
"loss": 1.0944,
|
| 2193 |
+
"step": 1555
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7202216066481995,
|
| 2197 |
+
"grad_norm": 3.3797718456778765,
|
| 2198 |
+
"learning_rate": 4.403320504472463e-06,
|
| 2199 |
+
"loss": 1.1426,
|
| 2200 |
+
"step": 1560
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7225300092336103,
|
| 2204 |
+
"grad_norm": 3.1832339015639826,
|
| 2205 |
+
"learning_rate": 4.336712560930891e-06,
|
| 2206 |
+
"loss": 1.1223,
|
| 2207 |
+
"step": 1565
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7248384118190212,
|
| 2211 |
+
"grad_norm": 3.40273969921815,
|
| 2212 |
+
"learning_rate": 4.270472477188755e-06,
|
| 2213 |
+
"loss": 1.1151,
|
| 2214 |
+
"step": 1570
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7271468144044322,
|
| 2218 |
+
"grad_norm": 3.32953363908172,
|
| 2219 |
+
"learning_rate": 4.204604555881967e-06,
|
| 2220 |
+
"loss": 1.1055,
|
| 2221 |
+
"step": 1575
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7294552169898431,
|
| 2225 |
+
"grad_norm": 3.363759228103856,
|
| 2226 |
+
"learning_rate": 4.139113075472565e-06,
|
| 2227 |
+
"loss": 1.15,
|
| 2228 |
+
"step": 1580
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7317636195752539,
|
| 2232 |
+
"grad_norm": 3.5692625205390214,
|
| 2233 |
+
"learning_rate": 4.074002289970801e-06,
|
| 2234 |
+
"loss": 1.1249,
|
| 2235 |
+
"step": 1585
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7340720221606648,
|
| 2239 |
+
"grad_norm": 3.6298117912857895,
|
| 2240 |
+
"learning_rate": 4.009276428658836e-06,
|
| 2241 |
+
"loss": 1.0911,
|
| 2242 |
+
"step": 1590
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7363804247460757,
|
| 2246 |
+
"grad_norm": 3.501911680130801,
|
| 2247 |
+
"learning_rate": 3.944939695816005e-06,
|
| 2248 |
+
"loss": 1.0591,
|
| 2249 |
+
"step": 1595
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7386888273314867,
|
| 2253 |
+
"grad_norm": 3.314254645913856,
|
| 2254 |
+
"learning_rate": 3.8809962704457375e-06,
|
| 2255 |
+
"loss": 1.122,
|
| 2256 |
+
"step": 1600
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7409972299168975,
|
| 2260 |
+
"grad_norm": 3.56145944415269,
|
| 2261 |
+
"learning_rate": 3.81745030600411e-06,
|
| 2262 |
+
"loss": 1.1036,
|
| 2263 |
+
"step": 1605
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7433056325023084,
|
| 2267 |
+
"grad_norm": 3.4910849192084235,
|
| 2268 |
+
"learning_rate": 3.75430593013006e-06,
|
| 2269 |
+
"loss": 1.1353,
|
| 2270 |
+
"step": 1610
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7456140350877193,
|
| 2274 |
+
"grad_norm": 3.325715619787326,
|
| 2275 |
+
"learning_rate": 3.6915672443772644e-06,
|
| 2276 |
+
"loss": 1.1538,
|
| 2277 |
+
"step": 1615
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.7479224376731302,
|
| 2281 |
+
"grad_norm": 3.5950013679874724,
|
| 2282 |
+
"learning_rate": 3.62923832394774e-06,
|
| 2283 |
+
"loss": 1.0909,
|
| 2284 |
+
"step": 1620
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7502308402585411,
|
| 2288 |
+
"grad_norm": 3.1524005532212334,
|
| 2289 |
+
"learning_rate": 3.56732321742712e-06,
|
| 2290 |
+
"loss": 1.1125,
|
| 2291 |
+
"step": 1625
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.752539242843952,
|
| 2295 |
+
"grad_norm": 3.6760451234626124,
|
| 2296 |
+
"learning_rate": 3.5058259465216828e-06,
|
| 2297 |
+
"loss": 1.1039,
|
| 2298 |
+
"step": 1630
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7548476454293629,
|
| 2302 |
+
"grad_norm": 3.341546948891595,
|
| 2303 |
+
"learning_rate": 3.444750505797123e-06,
|
| 2304 |
+
"loss": 1.0531,
|
| 2305 |
+
"step": 1635
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7571560480147738,
|
| 2309 |
+
"grad_norm": 3.35627649123262,
|
| 2310 |
+
"learning_rate": 3.384100862419096e-06,
|
| 2311 |
+
"loss": 1.0931,
|
| 2312 |
+
"step": 1640
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7594644506001846,
|
| 2316 |
+
"grad_norm": 3.6221419833131527,
|
| 2317 |
+
"learning_rate": 3.3238809558955054e-06,
|
| 2318 |
+
"loss": 1.0797,
|
| 2319 |
+
"step": 1645
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7617728531855956,
|
| 2323 |
+
"grad_norm": 3.3487671296828267,
|
| 2324 |
+
"learning_rate": 3.2640946978206266e-06,
|
| 2325 |
+
"loss": 1.0812,
|
| 2326 |
+
"step": 1650
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7640812557710065,
|
| 2330 |
+
"grad_norm": 3.441031645390376,
|
| 2331 |
+
"learning_rate": 3.2047459716210306e-06,
|
| 2332 |
+
"loss": 1.1155,
|
| 2333 |
+
"step": 1655
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7663896583564174,
|
| 2337 |
+
"grad_norm": 3.4825057106301096,
|
| 2338 |
+
"learning_rate": 3.145838632303325e-06,
|
| 2339 |
+
"loss": 1.096,
|
| 2340 |
+
"step": 1660
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.7686980609418282,
|
| 2344 |
+
"grad_norm": 3.4525699686491875,
|
| 2345 |
+
"learning_rate": 3.087376506203763e-06,
|
| 2346 |
+
"loss": 1.145,
|
| 2347 |
+
"step": 1665
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.7710064635272391,
|
| 2351 |
+
"grad_norm": 3.2639030957505715,
|
| 2352 |
+
"learning_rate": 3.0293633907396903e-06,
|
| 2353 |
+
"loss": 1.0711,
|
| 2354 |
+
"step": 1670
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7733148661126501,
|
| 2358 |
+
"grad_norm": 3.247147491878351,
|
| 2359 |
+
"learning_rate": 2.971803054162903e-06,
|
| 2360 |
+
"loss": 1.0367,
|
| 2361 |
+
"step": 1675
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.775623268698061,
|
| 2365 |
+
"grad_norm": 3.3628039668359824,
|
| 2366 |
+
"learning_rate": 2.914699235314855e-06,
|
| 2367 |
+
"loss": 1.1311,
|
| 2368 |
+
"step": 1680
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7779316712834718,
|
| 2372 |
+
"grad_norm": 3.294560766749018,
|
| 2373 |
+
"learning_rate": 2.858055643383818e-06,
|
| 2374 |
+
"loss": 1.1303,
|
| 2375 |
+
"step": 1685
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7802400738688827,
|
| 2379 |
+
"grad_norm": 3.252460881051861,
|
| 2380 |
+
"learning_rate": 2.8018759576639478e-06,
|
| 2381 |
+
"loss": 1.0894,
|
| 2382 |
+
"step": 1690
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7825484764542936,
|
| 2386 |
+
"grad_norm": 3.6541818791755083,
|
| 2387 |
+
"learning_rate": 2.7461638273162895e-06,
|
| 2388 |
+
"loss": 1.1416,
|
| 2389 |
+
"step": 1695
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7848568790397045,
|
| 2393 |
+
"grad_norm": 3.3018114290440286,
|
| 2394 |
+
"learning_rate": 2.6909228711317526e-06,
|
| 2395 |
+
"loss": 1.0898,
|
| 2396 |
+
"step": 1700
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7871652816251155,
|
| 2400 |
+
"grad_norm": 3.5110479717681704,
|
| 2401 |
+
"learning_rate": 2.6361566772960466e-06,
|
| 2402 |
+
"loss": 1.0887,
|
| 2403 |
+
"step": 1705
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7894736842105263,
|
| 2407 |
+
"grad_norm": 3.469571849173682,
|
| 2408 |
+
"learning_rate": 2.5818688031566132e-06,
|
| 2409 |
+
"loss": 1.0182,
|
| 2410 |
+
"step": 1710
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.7917820867959372,
|
| 2414 |
+
"grad_norm": 3.761287355693432,
|
| 2415 |
+
"learning_rate": 2.5280627749915544e-06,
|
| 2416 |
+
"loss": 1.1246,
|
| 2417 |
+
"step": 1715
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7940904893813481,
|
| 2421 |
+
"grad_norm": 3.7171990367681866,
|
| 2422 |
+
"learning_rate": 2.4747420877805905e-06,
|
| 2423 |
+
"loss": 1.1008,
|
| 2424 |
+
"step": 1720
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.796398891966759,
|
| 2428 |
+
"grad_norm": 3.583342537171837,
|
| 2429 |
+
"learning_rate": 2.421910204978033e-06,
|
| 2430 |
+
"loss": 1.092,
|
| 2431 |
+
"step": 1725
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7987072945521699,
|
| 2435 |
+
"grad_norm": 3.3105866570237343,
|
| 2436 |
+
"learning_rate": 2.369570558287819e-06,
|
| 2437 |
+
"loss": 1.0495,
|
| 2438 |
+
"step": 1730
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8010156971375808,
|
| 2442 |
+
"grad_norm": 3.453250654565143,
|
| 2443 |
+
"learning_rate": 2.3177265474406084e-06,
|
| 2444 |
+
"loss": 1.0952,
|
| 2445 |
+
"step": 1735
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8033240997229917,
|
| 2449 |
+
"grad_norm": 3.2111312681793294,
|
| 2450 |
+
"learning_rate": 2.2663815399729495e-06,
|
| 2451 |
+
"loss": 1.0756,
|
| 2452 |
+
"step": 1740
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.8056325023084026,
|
| 2456 |
+
"grad_norm": 3.398739502823191,
|
| 2457 |
+
"learning_rate": 2.215538871008538e-06,
|
| 2458 |
+
"loss": 1.0855,
|
| 2459 |
+
"step": 1745
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8079409048938134,
|
| 2463 |
+
"grad_norm": 3.4089573083048883,
|
| 2464 |
+
"learning_rate": 2.1652018430415923e-06,
|
| 2465 |
+
"loss": 1.0707,
|
| 2466 |
+
"step": 1750
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8102493074792244,
|
| 2470 |
+
"grad_norm": 3.7996382043873744,
|
| 2471 |
+
"learning_rate": 2.115373725722326e-06,
|
| 2472 |
+
"loss": 1.1419,
|
| 2473 |
+
"step": 1755
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8125577100646353,
|
| 2477 |
+
"grad_norm": 3.4303103622199203,
|
| 2478 |
+
"learning_rate": 2.066057755644587e-06,
|
| 2479 |
+
"loss": 1.1101,
|
| 2480 |
+
"step": 1760
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8148661126500462,
|
| 2484 |
+
"grad_norm": 3.3758394994097363,
|
| 2485 |
+
"learning_rate": 2.0172571361356007e-06,
|
| 2486 |
+
"loss": 1.0975,
|
| 2487 |
+
"step": 1765
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.817174515235457,
|
| 2491 |
+
"grad_norm": 3.2901551940425673,
|
| 2492 |
+
"learning_rate": 1.9689750370479134e-06,
|
| 2493 |
+
"loss": 1.0797,
|
| 2494 |
+
"step": 1770
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8194829178208679,
|
| 2498 |
+
"grad_norm": 3.661068632899665,
|
| 2499 |
+
"learning_rate": 1.921214594553488e-06,
|
| 2500 |
+
"loss": 1.1287,
|
| 2501 |
+
"step": 1775
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8217913204062789,
|
| 2505 |
+
"grad_norm": 3.5442080312978415,
|
| 2506 |
+
"learning_rate": 1.8739789109399954e-06,
|
| 2507 |
+
"loss": 1.1514,
|
| 2508 |
+
"step": 1780
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8240997229916898,
|
| 2512 |
+
"grad_norm": 3.3534741257777325,
|
| 2513 |
+
"learning_rate": 1.8272710544093019e-06,
|
| 2514 |
+
"loss": 1.0824,
|
| 2515 |
+
"step": 1785
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8264081255771006,
|
| 2519 |
+
"grad_norm": 3.570055818522298,
|
| 2520 |
+
"learning_rate": 1.7810940588781811e-06,
|
| 2521 |
+
"loss": 1.1313,
|
| 2522 |
+
"step": 1790
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8287165281625115,
|
| 2526 |
+
"grad_norm": 3.3907592881825352,
|
| 2527 |
+
"learning_rate": 1.7354509237812334e-06,
|
| 2528 |
+
"loss": 1.0458,
|
| 2529 |
+
"step": 1795
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8310249307479224,
|
| 2533 |
+
"grad_norm": 3.7660635086416794,
|
| 2534 |
+
"learning_rate": 1.690344613876066e-06,
|
| 2535 |
+
"loss": 1.109,
|
| 2536 |
+
"step": 1800
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8333333333333334,
|
| 2540 |
+
"grad_norm": 20.624336407323348,
|
| 2541 |
+
"learning_rate": 1.64577805905072e-06,
|
| 2542 |
+
"loss": 1.0872,
|
| 2543 |
+
"step": 1805
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8356417359187442,
|
| 2547 |
+
"grad_norm": 3.38434013035599,
|
| 2548 |
+
"learning_rate": 1.601754154133347e-06,
|
| 2549 |
+
"loss": 1.0943,
|
| 2550 |
+
"step": 1810
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8379501385041551,
|
| 2554 |
+
"grad_norm": 3.3282071197431318,
|
| 2555 |
+
"learning_rate": 1.558275758704183e-06,
|
| 2556 |
+
"loss": 1.0983,
|
| 2557 |
+
"step": 1815
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.840258541089566,
|
| 2561 |
+
"grad_norm": 3.4156960745203286,
|
| 2562 |
+
"learning_rate": 1.5153456969098013e-06,
|
| 2563 |
+
"loss": 1.0381,
|
| 2564 |
+
"step": 1820
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8425669436749769,
|
| 2568 |
+
"grad_norm": 3.3418973703656274,
|
| 2569 |
+
"learning_rate": 1.4729667572796735e-06,
|
| 2570 |
+
"loss": 1.1452,
|
| 2571 |
+
"step": 1825
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8448753462603878,
|
| 2575 |
+
"grad_norm": 3.333897377962453,
|
| 2576 |
+
"learning_rate": 1.431141692545036e-06,
|
| 2577 |
+
"loss": 1.1076,
|
| 2578 |
+
"step": 1830
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8471837488457987,
|
| 2582 |
+
"grad_norm": 3.402941306050666,
|
| 2583 |
+
"learning_rate": 1.389873219460085e-06,
|
| 2584 |
+
"loss": 1.0869,
|
| 2585 |
+
"step": 1835
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.8494921514312096,
|
| 2589 |
+
"grad_norm": 3.3313186519496423,
|
| 2590 |
+
"learning_rate": 1.349164018625513e-06,
|
| 2591 |
+
"loss": 1.0765,
|
| 2592 |
+
"step": 1840
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8518005540166205,
|
| 2596 |
+
"grad_norm": 3.6011720414080566,
|
| 2597 |
+
"learning_rate": 1.3090167343143911e-06,
|
| 2598 |
+
"loss": 1.0846,
|
| 2599 |
+
"step": 1845
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8541089566020313,
|
| 2603 |
+
"grad_norm": 3.629326020817196,
|
| 2604 |
+
"learning_rate": 1.2694339743004037e-06,
|
| 2605 |
+
"loss": 1.1088,
|
| 2606 |
+
"step": 1850
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8564173591874423,
|
| 2610 |
+
"grad_norm": 3.6305906598709767,
|
| 2611 |
+
"learning_rate": 1.2304183096884626e-06,
|
| 2612 |
+
"loss": 1.0875,
|
| 2613 |
+
"step": 1855
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8587257617728532,
|
| 2617 |
+
"grad_norm": 3.35865168543221,
|
| 2618 |
+
"learning_rate": 1.1919722747477024e-06,
|
| 2619 |
+
"loss": 1.1143,
|
| 2620 |
+
"step": 1860
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8610341643582641,
|
| 2624 |
+
"grad_norm": 3.3889339992199177,
|
| 2625 |
+
"learning_rate": 1.1540983667468686e-06,
|
| 2626 |
+
"loss": 1.0916,
|
| 2627 |
+
"step": 1865
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.863342566943675,
|
| 2631 |
+
"grad_norm": 3.3133014347890324,
|
| 2632 |
+
"learning_rate": 1.1167990457920985e-06,
|
| 2633 |
+
"loss": 1.0877,
|
| 2634 |
+
"step": 1870
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8656509695290858,
|
| 2638 |
+
"grad_norm": 3.415023896862017,
|
| 2639 |
+
"learning_rate": 1.0800767346671347e-06,
|
| 2640 |
+
"loss": 1.0284,
|
| 2641 |
+
"step": 1875
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8679593721144968,
|
| 2645 |
+
"grad_norm": 3.322962975732958,
|
| 2646 |
+
"learning_rate": 1.043933818675944e-06,
|
| 2647 |
+
"loss": 1.0782,
|
| 2648 |
+
"step": 1880
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8702677746999077,
|
| 2652 |
+
"grad_norm": 3.583896655771928,
|
| 2653 |
+
"learning_rate": 1.008372645487785e-06,
|
| 2654 |
+
"loss": 1.08,
|
| 2655 |
+
"step": 1885
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8725761772853186,
|
| 2659 |
+
"grad_norm": 3.3057678718948726,
|
| 2660 |
+
"learning_rate": 9.733955249847183e-07,
|
| 2661 |
+
"loss": 1.1034,
|
| 2662 |
+
"step": 1890
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8748845798707294,
|
| 2666 |
+
"grad_norm": 3.4387092657320997,
|
| 2667 |
+
"learning_rate": 9.390047291115567e-07,
|
| 2668 |
+
"loss": 1.0915,
|
| 2669 |
+
"step": 1895
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8771929824561403,
|
| 2673 |
+
"grad_norm": 3.8029482282950324,
|
| 2674 |
+
"learning_rate": 9.052024917282987e-07,
|
| 2675 |
+
"loss": 1.057,
|
| 2676 |
+
"step": 1900
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8795013850415513,
|
| 2680 |
+
"grad_norm": 3.3990790831971465,
|
| 2681 |
+
"learning_rate": 8.719910084650262e-07,
|
| 2682 |
+
"loss": 1.0725,
|
| 2683 |
+
"step": 1905
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8818097876269622,
|
| 2687 |
+
"grad_norm": 3.262416726762208,
|
| 2688 |
+
"learning_rate": 8.393724365792866e-07,
|
| 2689 |
+
"loss": 1.1028,
|
| 2690 |
+
"step": 1910
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.884118190212373,
|
| 2694 |
+
"grad_norm": 3.551691283783414,
|
| 2695 |
+
"learning_rate": 8.073488948159691e-07,
|
| 2696 |
+
"loss": 1.0546,
|
| 2697 |
+
"step": 1915
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.8864265927977839,
|
| 2701 |
+
"grad_norm": 3.5211563130144197,
|
| 2702 |
+
"learning_rate": 7.759224632696793e-07,
|
| 2703 |
+
"loss": 1.1024,
|
| 2704 |
+
"step": 1920
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.8887349953831948,
|
| 2708 |
+
"grad_norm": 3.5958803804208976,
|
| 2709 |
+
"learning_rate": 7.450951832496233e-07,
|
| 2710 |
+
"loss": 1.0698,
|
| 2711 |
+
"step": 1925
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.8910433979686058,
|
| 2715 |
+
"grad_norm": 4.107811963680795,
|
| 2716 |
+
"learning_rate": 7.148690571470251e-07,
|
| 2717 |
+
"loss": 1.0613,
|
| 2718 |
+
"step": 1930
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8933518005540166,
|
| 2722 |
+
"grad_norm": 3.6280688174940416,
|
| 2723 |
+
"learning_rate": 6.852460483050494e-07,
|
| 2724 |
+
"loss": 1.0987,
|
| 2725 |
+
"step": 1935
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8956602031394275,
|
| 2729 |
+
"grad_norm": 3.4197153407779055,
|
| 2730 |
+
"learning_rate": 6.562280808912768e-07,
|
| 2731 |
+
"loss": 1.081,
|
| 2732 |
+
"step": 1940
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8979686057248384,
|
| 2736 |
+
"grad_norm": 3.3975321682078494,
|
| 2737 |
+
"learning_rate": 6.278170397727179e-07,
|
| 2738 |
+
"loss": 1.0881,
|
| 2739 |
+
"step": 1945
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.9002770083102493,
|
| 2743 |
+
"grad_norm": 3.385824657440924,
|
| 2744 |
+
"learning_rate": 6.000147703933845e-07,
|
| 2745 |
+
"loss": 1.0725,
|
| 2746 |
+
"step": 1950
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.9025854108956602,
|
| 2750 |
+
"grad_norm": 3.733106691108023,
|
| 2751 |
+
"learning_rate": 5.728230786544153e-07,
|
| 2752 |
+
"loss": 1.0886,
|
| 2753 |
+
"step": 1955
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.9048938134810711,
|
| 2757 |
+
"grad_norm": 3.3831515124529288,
|
| 2758 |
+
"learning_rate": 5.46243730796776e-07,
|
| 2759 |
+
"loss": 1.0854,
|
| 2760 |
+
"step": 1960
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.907202216066482,
|
| 2764 |
+
"grad_norm": 3.4106013139907065,
|
| 2765 |
+
"learning_rate": 5.202784532865302e-07,
|
| 2766 |
+
"loss": 1.114,
|
| 2767 |
+
"step": 1965
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.9095106186518929,
|
| 2771 |
+
"grad_norm": 3.130011381325973,
|
| 2772 |
+
"learning_rate": 4.949289327026952e-07,
|
| 2773 |
+
"loss": 1.0873,
|
| 2774 |
+
"step": 1970
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.9118190212373037,
|
| 2778 |
+
"grad_norm": 3.3600219468750394,
|
| 2779 |
+
"learning_rate": 4.7019681562769816e-07,
|
| 2780 |
+
"loss": 1.0689,
|
| 2781 |
+
"step": 1975
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9141274238227147,
|
| 2785 |
+
"grad_norm": 3.379655670825615,
|
| 2786 |
+
"learning_rate": 4.460837085404113e-07,
|
| 2787 |
+
"loss": 1.0874,
|
| 2788 |
+
"step": 1980
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9164358264081256,
|
| 2792 |
+
"grad_norm": 3.324809563310868,
|
| 2793 |
+
"learning_rate": 4.225911777118097e-07,
|
| 2794 |
+
"loss": 1.0894,
|
| 2795 |
+
"step": 1985
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9187442289935365,
|
| 2799 |
+
"grad_norm": 3.4668181744618196,
|
| 2800 |
+
"learning_rate": 3.9972074910323066e-07,
|
| 2801 |
+
"loss": 1.0896,
|
| 2802 |
+
"step": 1990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9210526315789473,
|
| 2806 |
+
"grad_norm": 3.4175120046363276,
|
| 2807 |
+
"learning_rate": 3.7747390826725736e-07,
|
| 2808 |
+
"loss": 1.0608,
|
| 2809 |
+
"step": 1995
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9233610341643582,
|
| 2813 |
+
"grad_norm": 3.365932789028912,
|
| 2814 |
+
"learning_rate": 3.5585210025122166e-07,
|
| 2815 |
+
"loss": 1.0465,
|
| 2816 |
+
"step": 2000
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9256694367497692,
|
| 2820 |
+
"grad_norm": 3.3721429412301442,
|
| 2821 |
+
"learning_rate": 3.3485672950334447e-07,
|
| 2822 |
+
"loss": 1.0782,
|
| 2823 |
+
"step": 2005
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9279778393351801,
|
| 2827 |
+
"grad_norm": 3.402893452692765,
|
| 2828 |
+
"learning_rate": 3.1448915978150365e-07,
|
| 2829 |
+
"loss": 1.0575,
|
| 2830 |
+
"step": 2010
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.930286241920591,
|
| 2834 |
+
"grad_norm": 3.3246351042614606,
|
| 2835 |
+
"learning_rate": 2.947507140646588e-07,
|
| 2836 |
+
"loss": 1.093,
|
| 2837 |
+
"step": 2015
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9325946445060018,
|
| 2841 |
+
"grad_norm": 3.42392243323848,
|
| 2842 |
+
"learning_rate": 2.756426744669105e-07,
|
| 2843 |
+
"loss": 1.0709,
|
| 2844 |
+
"step": 2020
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9349030470914127,
|
| 2848 |
+
"grad_norm": 3.3870385964627565,
|
| 2849 |
+
"learning_rate": 2.57166282154222e-07,
|
| 2850 |
+
"loss": 1.0944,
|
| 2851 |
+
"step": 2025
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.9372114496768237,
|
| 2855 |
+
"grad_norm": 3.4345800654530128,
|
| 2856 |
+
"learning_rate": 2.393227372638018e-07,
|
| 2857 |
+
"loss": 1.0829,
|
| 2858 |
+
"step": 2030
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.9395198522622346,
|
| 2862 |
+
"grad_norm": 3.2304527099741094,
|
| 2863 |
+
"learning_rate": 2.221131988261438e-07,
|
| 2864 |
+
"loss": 1.0663,
|
| 2865 |
+
"step": 2035
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9418282548476454,
|
| 2869 |
+
"grad_norm": 3.4212248154000324,
|
| 2870 |
+
"learning_rate": 2.055387846897472e-07,
|
| 2871 |
+
"loss": 1.0608,
|
| 2872 |
+
"step": 2040
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9441366574330563,
|
| 2876 |
+
"grad_norm": 3.3424495231710356,
|
| 2877 |
+
"learning_rate": 1.8960057144850163e-07,
|
| 2878 |
+
"loss": 1.0513,
|
| 2879 |
+
"step": 2045
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9464450600184672,
|
| 2883 |
+
"grad_norm": 8.42913586604929,
|
| 2884 |
+
"learning_rate": 1.742995943717607e-07,
|
| 2885 |
+
"loss": 1.0698,
|
| 2886 |
+
"step": 2050
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.9487534626038782,
|
| 2890 |
+
"grad_norm": 4.03605470816158,
|
| 2891 |
+
"learning_rate": 1.5963684733709462e-07,
|
| 2892 |
+
"loss": 1.0787,
|
| 2893 |
+
"step": 2055
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.951061865189289,
|
| 2897 |
+
"grad_norm": 3.572766321551919,
|
| 2898 |
+
"learning_rate": 1.4561328276573415e-07,
|
| 2899 |
+
"loss": 1.0625,
|
| 2900 |
+
"step": 2060
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9533702677746999,
|
| 2904 |
+
"grad_norm": 3.213406555168112,
|
| 2905 |
+
"learning_rate": 1.3222981156070126e-07,
|
| 2906 |
+
"loss": 1.0861,
|
| 2907 |
+
"step": 2065
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9556786703601108,
|
| 2911 |
+
"grad_norm": 3.216022210724082,
|
| 2912 |
+
"learning_rate": 1.1948730304764622e-07,
|
| 2913 |
+
"loss": 1.0572,
|
| 2914 |
+
"step": 2070
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.9579870729455217,
|
| 2918 |
+
"grad_norm": 3.8142801195990237,
|
| 2919 |
+
"learning_rate": 1.073865849183786e-07,
|
| 2920 |
+
"loss": 1.1151,
|
| 2921 |
+
"step": 2075
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9602954755309326,
|
| 2925 |
+
"grad_norm": 3.2011503381896413,
|
| 2926 |
+
"learning_rate": 9.592844317710238e-08,
|
| 2927 |
+
"loss": 1.0585,
|
| 2928 |
+
"step": 2080
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9626038781163435,
|
| 2932 |
+
"grad_norm": 3.3780038857652226,
|
| 2933 |
+
"learning_rate": 8.511362208936447e-08,
|
| 2934 |
+
"loss": 1.0591,
|
| 2935 |
+
"step": 2085
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9649122807017544,
|
| 2939 |
+
"grad_norm": 3.3212612452494295,
|
| 2940 |
+
"learning_rate": 7.494282413371135e-08,
|
| 2941 |
+
"loss": 1.0787,
|
| 2942 |
+
"step": 2090
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9672206832871653,
|
| 2946 |
+
"grad_norm": 3.797857330316498,
|
| 2947 |
+
"learning_rate": 6.541670995605321e-08,
|
| 2948 |
+
"loss": 1.0859,
|
| 2949 |
+
"step": 2095
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.9695290858725761,
|
| 2953 |
+
"grad_norm": 3.153773745189338,
|
| 2954 |
+
"learning_rate": 5.653589832675943e-08,
|
| 2955 |
+
"loss": 1.0983,
|
| 2956 |
+
"step": 2100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9718374884579871,
|
| 2960 |
+
"grad_norm": 3.4652822167549906,
|
| 2961 |
+
"learning_rate": 4.830096610045854e-08,
|
| 2962 |
+
"loss": 1.0713,
|
| 2963 |
+
"step": 2105
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.974145891043398,
|
| 2967 |
+
"grad_norm": 3.6601967632905796,
|
| 2968 |
+
"learning_rate": 4.071244817857589e-08,
|
| 2969 |
+
"loss": 1.1118,
|
| 2970 |
+
"step": 2110
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9764542936288089,
|
| 2974 |
+
"grad_norm": 3.135385406063897,
|
| 2975 |
+
"learning_rate": 3.3770837474584874e-08,
|
| 2976 |
+
"loss": 1.072,
|
| 2977 |
+
"step": 2115
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9787626962142197,
|
| 2981 |
+
"grad_norm": 3.4884714571677784,
|
| 2982 |
+
"learning_rate": 2.747658488199023e-08,
|
| 2983 |
+
"loss": 1.0738,
|
| 2984 |
+
"step": 2120
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9810710987996306,
|
| 2988 |
+
"grad_norm": 3.7803263448925706,
|
| 2989 |
+
"learning_rate": 2.1830099245040427e-08,
|
| 2990 |
+
"loss": 1.0549,
|
| 2991 |
+
"step": 2125
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9833795013850416,
|
| 2995 |
+
"grad_norm": 3.3045019552603585,
|
| 2996 |
+
"learning_rate": 1.683174733216997e-08,
|
| 2997 |
+
"loss": 1.1129,
|
| 2998 |
+
"step": 2130
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9856879039704525,
|
| 3002 |
+
"grad_norm": 3.2212668180182784,
|
| 3003 |
+
"learning_rate": 1.248185381217848e-08,
|
| 3004 |
+
"loss": 1.0777,
|
| 3005 |
+
"step": 2135
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.9879963065558633,
|
| 3009 |
+
"grad_norm": 3.324768260260177,
|
| 3010 |
+
"learning_rate": 8.780701233139789e-09,
|
| 3011 |
+
"loss": 1.0503,
|
| 3012 |
+
"step": 2140
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.9903047091412742,
|
| 3016 |
+
"grad_norm": 3.214869100486745,
|
| 3017 |
+
"learning_rate": 5.728530004051047e-09,
|
| 3018 |
+
"loss": 1.0367,
|
| 3019 |
+
"step": 2145
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.9926131117266851,
|
| 3023 |
+
"grad_norm": 3.3583215428853666,
|
| 3024 |
+
"learning_rate": 3.325538379211901e-09,
|
| 3025 |
+
"loss": 1.0554,
|
| 3026 |
+
"step": 2150
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.9949215143120961,
|
| 3030 |
+
"grad_norm": 4.075445923312751,
|
| 3031 |
+
"learning_rate": 1.5718824453525572e-09,
|
| 3032 |
+
"loss": 1.1222,
|
| 3033 |
+
"step": 2155
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.997229916897507,
|
| 3037 |
+
"grad_norm": 3.3599147688364903,
|
| 3038 |
+
"learning_rate": 4.676761114941197e-10,
|
| 3039 |
+
"loss": 1.0646,
|
| 3040 |
+
"step": 2160
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9995383194829178,
|
| 3044 |
+
"grad_norm": 3.4496692841727543,
|
| 3045 |
+
"learning_rate": 1.2991101545622998e-11,
|
| 3046 |
+
"loss": 1.1038,
|
| 3047 |
+
"step": 2165
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 1.0,
|
| 3051 |
+
"eval_loss": 1.1177629232406616,
|
| 3052 |
+
"eval_runtime": 1154.8442,
|
| 3053 |
+
"eval_samples_per_second": 26.579,
|
| 3054 |
+
"eval_steps_per_second": 0.831,
|
| 3055 |
+
"step": 2166
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 1.0,
|
| 3059 |
+
"step": 2166,
|
| 3060 |
+
"total_flos": 113379083550720.0,
|
| 3061 |
+
"train_loss": 1.171416565762112,
|
| 3062 |
+
"train_runtime": 11018.7418,
|
| 3063 |
+
"train_samples_per_second": 6.29,
|
| 3064 |
+
"train_steps_per_second": 0.197
|
| 3065 |
+
}
|
| 3066 |
+
],
|
| 3067 |
+
"logging_steps": 5,
|
| 3068 |
+
"max_steps": 2166,
|
| 3069 |
+
"num_input_tokens_seen": 0,
|
| 3070 |
+
"num_train_epochs": 1,
|
| 3071 |
+
"save_steps": 500,
|
| 3072 |
+
"stateful_callbacks": {
|
| 3073 |
+
"TrainerControl": {
|
| 3074 |
+
"args": {
|
| 3075 |
+
"should_epoch_stop": false,
|
| 3076 |
+
"should_evaluate": false,
|
| 3077 |
+
"should_log": false,
|
| 3078 |
+
"should_save": true,
|
| 3079 |
+
"should_training_stop": true
|
| 3080 |
+
},
|
| 3081 |
+
"attributes": {}
|
| 3082 |
+
}
|
| 3083 |
+
},
|
| 3084 |
+
"total_flos": 113379083550720.0,
|
| 3085 |
+
"train_batch_size": 8,
|
| 3086 |
+
"trial_name": null,
|
| 3087 |
+
"trial_params": null
|
| 3088 |
+
}
|