Model save
Browse files- README.md +67 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +1050 -0
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: google/gemma-7b
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: gemma-7b-slic-basic-5e-5-05-v5
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- orpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for gemma-7b-slic-basic-5e-5-05-v5
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b).
|
| 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="silviasapora/gemma-7b-slic-basic-5e-5-05-v5", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/silvias/huggingface/runs/qqrikeob)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with ORPO, a method introduced in [ORPO: Monolithic Preference Optimization without Reference Model](https://huggingface.co/papers/2403.07691).
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.13.0
|
| 38 |
+
- Transformers: 4.46.1
|
| 39 |
+
- Pytorch: 2.4.0
|
| 40 |
+
- Datasets: 3.1.0
|
| 41 |
+
- Tokenizers: 0.20.1
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite ORPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{hong2024orpo,
|
| 49 |
+
title = {{ORPO: Monolithic Preference Optimization without Reference Model}},
|
| 50 |
+
author = {Jiwoo Hong and Noah Lee and James Thorne},
|
| 51 |
+
year = 2024,
|
| 52 |
+
eprint = {arXiv:2403.07691}
|
| 53 |
+
}
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
Cite TRL as:
|
| 57 |
+
|
| 58 |
+
```bibtex
|
| 59 |
+
@misc{vonwerra2022trl,
|
| 60 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 61 |
+
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},
|
| 62 |
+
year = 2020,
|
| 63 |
+
journal = {GitHub repository},
|
| 64 |
+
publisher = {GitHub},
|
| 65 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 66 |
+
}
|
| 67 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.986666666666667,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 25.709435809604706,
|
| 5 |
+
"train_runtime": 7268.7111,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.786,
|
| 8 |
+
"train_steps_per_second": 0.043
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.986666666666667,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 25.709435809604706,
|
| 5 |
+
"train_runtime": 7268.7111,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.786,
|
| 8 |
+
"train_steps_per_second": 0.043
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1050 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.986666666666667,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 315,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.047407407407407405,
|
| 13 |
+
"grad_norm": 1072.0,
|
| 14 |
+
"learning_rate": 7.8125e-06,
|
| 15 |
+
"log_odds_chosen": -1.060240626335144,
|
| 16 |
+
"log_odds_ratio": -11.007429122924805,
|
| 17 |
+
"logps/chosen": -22.712669372558594,
|
| 18 |
+
"logps/rejected": -21.652647018432617,
|
| 19 |
+
"loss": 455.4558,
|
| 20 |
+
"nll_loss": 9.838397026062012,
|
| 21 |
+
"rewards/accuracies": 0.4781250059604645,
|
| 22 |
+
"rewards/chosen": -11.356334686279297,
|
| 23 |
+
"rewards/margins": -0.530011773109436,
|
| 24 |
+
"rewards/rejected": -10.826323509216309,
|
| 25 |
+
"step": 5
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.09481481481481481,
|
| 29 |
+
"grad_norm": 592.0,
|
| 30 |
+
"learning_rate": 1.5625e-05,
|
| 31 |
+
"log_odds_chosen": -3.1204988956451416,
|
| 32 |
+
"log_odds_ratio": -11.043037414550781,
|
| 33 |
+
"logps/chosen": -21.994853973388672,
|
| 34 |
+
"logps/rejected": -18.874696731567383,
|
| 35 |
+
"loss": 424.5746,
|
| 36 |
+
"nll_loss": 8.752840042114258,
|
| 37 |
+
"rewards/accuracies": 0.4437499940395355,
|
| 38 |
+
"rewards/chosen": -10.997426986694336,
|
| 39 |
+
"rewards/margins": -1.5600786209106445,
|
| 40 |
+
"rewards/rejected": -9.437348365783691,
|
| 41 |
+
"step": 10
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.14222222222222222,
|
| 45 |
+
"grad_norm": 454.0,
|
| 46 |
+
"learning_rate": 2.34375e-05,
|
| 47 |
+
"log_odds_chosen": -1.2841793298721313,
|
| 48 |
+
"log_odds_ratio": -11.318960189819336,
|
| 49 |
+
"logps/chosen": -21.992443084716797,
|
| 50 |
+
"logps/rejected": -20.706798553466797,
|
| 51 |
+
"loss": 399.2196,
|
| 52 |
+
"nll_loss": 8.00178050994873,
|
| 53 |
+
"rewards/accuracies": 0.512499988079071,
|
| 54 |
+
"rewards/chosen": -10.996221542358398,
|
| 55 |
+
"rewards/margins": -0.6428214907646179,
|
| 56 |
+
"rewards/rejected": -10.353399276733398,
|
| 57 |
+
"step": 15
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.18962962962962962,
|
| 61 |
+
"grad_norm": 3008.0,
|
| 62 |
+
"learning_rate": 3.125e-05,
|
| 63 |
+
"log_odds_chosen": -4.260960578918457,
|
| 64 |
+
"log_odds_ratio": -9.433350563049316,
|
| 65 |
+
"logps/chosen": -17.13646125793457,
|
| 66 |
+
"logps/rejected": -12.87755298614502,
|
| 67 |
+
"loss": 320.5827,
|
| 68 |
+
"nll_loss": 6.233663558959961,
|
| 69 |
+
"rewards/accuracies": 0.44062501192092896,
|
| 70 |
+
"rewards/chosen": -8.568230628967285,
|
| 71 |
+
"rewards/margins": -2.129453659057617,
|
| 72 |
+
"rewards/rejected": -6.43877649307251,
|
| 73 |
+
"step": 20
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.23703703703703705,
|
| 77 |
+
"grad_norm": 296.0,
|
| 78 |
+
"learning_rate": 3.90625e-05,
|
| 79 |
+
"log_odds_chosen": -0.10830163955688477,
|
| 80 |
+
"log_odds_ratio": -1.7876460552215576,
|
| 81 |
+
"logps/chosen": -3.861504316329956,
|
| 82 |
+
"logps/rejected": -3.7353768348693848,
|
| 83 |
+
"loss": 92.4915,
|
| 84 |
+
"nll_loss": 2.6942696571350098,
|
| 85 |
+
"rewards/accuracies": 0.5531250238418579,
|
| 86 |
+
"rewards/chosen": -1.930752158164978,
|
| 87 |
+
"rewards/margins": -0.06306387484073639,
|
| 88 |
+
"rewards/rejected": -1.8676884174346924,
|
| 89 |
+
"step": 25
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.28444444444444444,
|
| 93 |
+
"grad_norm": 189.0,
|
| 94 |
+
"learning_rate": 4.6875e-05,
|
| 95 |
+
"log_odds_chosen": 0.08914291858673096,
|
| 96 |
+
"log_odds_ratio": -0.9090879559516907,
|
| 97 |
+
"logps/chosen": -1.9305442571640015,
|
| 98 |
+
"logps/rejected": -2.0069050788879395,
|
| 99 |
+
"loss": 65.6584,
|
| 100 |
+
"nll_loss": 2.0820810794830322,
|
| 101 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 102 |
+
"rewards/chosen": -0.9652721285820007,
|
| 103 |
+
"rewards/margins": 0.03818041831254959,
|
| 104 |
+
"rewards/rejected": -1.0034525394439697,
|
| 105 |
+
"step": 30
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33185185185185184,
|
| 109 |
+
"grad_norm": 247.0,
|
| 110 |
+
"learning_rate": 4.998613757348784e-05,
|
| 111 |
+
"log_odds_chosen": 0.3332863748073578,
|
| 112 |
+
"log_odds_ratio": -0.8136388063430786,
|
| 113 |
+
"logps/chosen": -1.8473803997039795,
|
| 114 |
+
"logps/rejected": -2.1461009979248047,
|
| 115 |
+
"loss": 58.187,
|
| 116 |
+
"nll_loss": 1.9638898372650146,
|
| 117 |
+
"rewards/accuracies": 0.550000011920929,
|
| 118 |
+
"rewards/chosen": -0.9236901998519897,
|
| 119 |
+
"rewards/margins": 0.14936023950576782,
|
| 120 |
+
"rewards/rejected": -1.0730504989624023,
|
| 121 |
+
"step": 35
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.37925925925925924,
|
| 125 |
+
"grad_norm": 316.0,
|
| 126 |
+
"learning_rate": 4.990147841143462e-05,
|
| 127 |
+
"log_odds_chosen": 0.5912309885025024,
|
| 128 |
+
"log_odds_ratio": -0.6898080110549927,
|
| 129 |
+
"logps/chosen": -1.7279255390167236,
|
| 130 |
+
"logps/rejected": -2.248690128326416,
|
| 131 |
+
"loss": 53.4463,
|
| 132 |
+
"nll_loss": 1.926064133644104,
|
| 133 |
+
"rewards/accuracies": 0.65625,
|
| 134 |
+
"rewards/chosen": -0.8639627695083618,
|
| 135 |
+
"rewards/margins": 0.2603822350502014,
|
| 136 |
+
"rewards/rejected": -1.124345064163208,
|
| 137 |
+
"step": 40
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.4266666666666667,
|
| 141 |
+
"grad_norm": 436.0,
|
| 142 |
+
"learning_rate": 4.97401218720448e-05,
|
| 143 |
+
"log_odds_chosen": 1.5738911628723145,
|
| 144 |
+
"log_odds_ratio": -0.6931017637252808,
|
| 145 |
+
"logps/chosen": -1.9965721368789673,
|
| 146 |
+
"logps/rejected": -3.4913876056671143,
|
| 147 |
+
"loss": 42.7119,
|
| 148 |
+
"nll_loss": 1.9767124652862549,
|
| 149 |
+
"rewards/accuracies": 0.684374988079071,
|
| 150 |
+
"rewards/chosen": -0.9982860684394836,
|
| 151 |
+
"rewards/margins": 0.7474077939987183,
|
| 152 |
+
"rewards/rejected": -1.7456938028335571,
|
| 153 |
+
"step": 45
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.4740740740740741,
|
| 157 |
+
"grad_norm": 340.0,
|
| 158 |
+
"learning_rate": 4.9502564938797946e-05,
|
| 159 |
+
"log_odds_chosen": 0.7620365619659424,
|
| 160 |
+
"log_odds_ratio": -0.7459171414375305,
|
| 161 |
+
"logps/chosen": -1.7422816753387451,
|
| 162 |
+
"logps/rejected": -2.4515380859375,
|
| 163 |
+
"loss": 52.5226,
|
| 164 |
+
"nll_loss": 1.9401613473892212,
|
| 165 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 166 |
+
"rewards/chosen": -0.8711408376693726,
|
| 167 |
+
"rewards/margins": 0.3546280860900879,
|
| 168 |
+
"rewards/rejected": -1.22576904296875,
|
| 169 |
+
"step": 50
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.5214814814814814,
|
| 173 |
+
"grad_norm": 336.0,
|
| 174 |
+
"learning_rate": 4.918953929490768e-05,
|
| 175 |
+
"log_odds_chosen": 2.066755533218384,
|
| 176 |
+
"log_odds_ratio": -0.703250527381897,
|
| 177 |
+
"logps/chosen": -2.0667014122009277,
|
| 178 |
+
"logps/rejected": -4.0496721267700195,
|
| 179 |
+
"loss": 43.0197,
|
| 180 |
+
"nll_loss": 2.0524208545684814,
|
| 181 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 182 |
+
"rewards/chosen": -1.0333507061004639,
|
| 183 |
+
"rewards/margins": 0.9914852380752563,
|
| 184 |
+
"rewards/rejected": -2.0248360633850098,
|
| 185 |
+
"step": 55
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.5688888888888889,
|
| 189 |
+
"grad_norm": 310.0,
|
| 190 |
+
"learning_rate": 4.88020090697132e-05,
|
| 191 |
+
"log_odds_chosen": 1.32530677318573,
|
| 192 |
+
"log_odds_ratio": -0.7277625799179077,
|
| 193 |
+
"logps/chosen": -1.7378164529800415,
|
| 194 |
+
"logps/rejected": -2.996918201446533,
|
| 195 |
+
"loss": 46.4442,
|
| 196 |
+
"nll_loss": 1.8602434396743774,
|
| 197 |
+
"rewards/accuracies": 0.6656249761581421,
|
| 198 |
+
"rewards/chosen": -0.8689082264900208,
|
| 199 |
+
"rewards/margins": 0.6295510530471802,
|
| 200 |
+
"rewards/rejected": -1.4984591007232666,
|
| 201 |
+
"step": 60
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.6162962962962963,
|
| 205 |
+
"grad_norm": 418.0,
|
| 206 |
+
"learning_rate": 4.834116786912897e-05,
|
| 207 |
+
"log_odds_chosen": 2.172917604446411,
|
| 208 |
+
"log_odds_ratio": -0.7570880651473999,
|
| 209 |
+
"logps/chosen": -1.923484444618225,
|
| 210 |
+
"logps/rejected": -4.014767169952393,
|
| 211 |
+
"loss": 43.8154,
|
| 212 |
+
"nll_loss": 2.0151474475860596,
|
| 213 |
+
"rewards/accuracies": 0.671875,
|
| 214 |
+
"rewards/chosen": -0.9617422223091125,
|
| 215 |
+
"rewards/margins": 1.0456410646438599,
|
| 216 |
+
"rewards/rejected": -2.0073835849761963,
|
| 217 |
+
"step": 65
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.6637037037037037,
|
| 221 |
+
"grad_norm": 362.0,
|
| 222 |
+
"learning_rate": 4.7808435099299045e-05,
|
| 223 |
+
"log_odds_chosen": 3.7954392433166504,
|
| 224 |
+
"log_odds_ratio": -0.6524917483329773,
|
| 225 |
+
"logps/chosen": -2.0215845108032227,
|
| 226 |
+
"logps/rejected": -5.7059245109558105,
|
| 227 |
+
"loss": 31.4526,
|
| 228 |
+
"nll_loss": 2.0003814697265625,
|
| 229 |
+
"rewards/accuracies": 0.753125011920929,
|
| 230 |
+
"rewards/chosen": -1.0107922554016113,
|
| 231 |
+
"rewards/margins": 1.842170000076294,
|
| 232 |
+
"rewards/rejected": -2.8529622554779053,
|
| 233 |
+
"step": 70
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.7111111111111111,
|
| 237 |
+
"grad_norm": 644.0,
|
| 238 |
+
"learning_rate": 4.720545159477922e-05,
|
| 239 |
+
"log_odds_chosen": 2.579057216644287,
|
| 240 |
+
"log_odds_ratio": -0.9140008687973022,
|
| 241 |
+
"logps/chosen": -2.144348382949829,
|
| 242 |
+
"logps/rejected": -4.612305641174316,
|
| 243 |
+
"loss": 40.0912,
|
| 244 |
+
"nll_loss": 1.925865888595581,
|
| 245 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 246 |
+
"rewards/chosen": -1.0721741914749146,
|
| 247 |
+
"rewards/margins": 1.2339783906936646,
|
| 248 |
+
"rewards/rejected": -2.306152820587158,
|
| 249 |
+
"step": 75
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.7585185185185185,
|
| 253 |
+
"grad_norm": 474.0,
|
| 254 |
+
"learning_rate": 4.653407456471222e-05,
|
| 255 |
+
"log_odds_chosen": 2.602337121963501,
|
| 256 |
+
"log_odds_ratio": -0.8660507202148438,
|
| 257 |
+
"logps/chosen": -2.392124891281128,
|
| 258 |
+
"logps/rejected": -4.895805358886719,
|
| 259 |
+
"loss": 45.2894,
|
| 260 |
+
"nll_loss": 2.1908349990844727,
|
| 261 |
+
"rewards/accuracies": 0.6781250238418579,
|
| 262 |
+
"rewards/chosen": -1.196062445640564,
|
| 263 |
+
"rewards/margins": 1.2518401145935059,
|
| 264 |
+
"rewards/rejected": -2.4479026794433594,
|
| 265 |
+
"step": 80
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.8059259259259259,
|
| 269 |
+
"grad_norm": 268.0,
|
| 270 |
+
"learning_rate": 4.579637187256222e-05,
|
| 271 |
+
"log_odds_chosen": 1.2037420272827148,
|
| 272 |
+
"log_odds_ratio": -0.6474961042404175,
|
| 273 |
+
"logps/chosen": -1.5563756227493286,
|
| 274 |
+
"logps/rejected": -2.667034864425659,
|
| 275 |
+
"loss": 40.2868,
|
| 276 |
+
"nll_loss": 1.7489856481552124,
|
| 277 |
+
"rewards/accuracies": 0.65625,
|
| 278 |
+
"rewards/chosen": -0.7781878113746643,
|
| 279 |
+
"rewards/margins": 0.5553296208381653,
|
| 280 |
+
"rewards/rejected": -1.3335174322128296,
|
| 281 |
+
"step": 85
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.8533333333333334,
|
| 285 |
+
"grad_norm": 372.0,
|
| 286 |
+
"learning_rate": 4.499461566702685e-05,
|
| 287 |
+
"log_odds_chosen": 5.79607629776001,
|
| 288 |
+
"log_odds_ratio": -0.834893524646759,
|
| 289 |
+
"logps/chosen": -2.230257987976074,
|
| 290 |
+
"logps/rejected": -7.889917850494385,
|
| 291 |
+
"loss": 33.4453,
|
| 292 |
+
"nll_loss": 2.099855422973633,
|
| 293 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 294 |
+
"rewards/chosen": -1.115128993988037,
|
| 295 |
+
"rewards/margins": 2.8298299312591553,
|
| 296 |
+
"rewards/rejected": -3.9449589252471924,
|
| 297 |
+
"step": 90
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.9007407407407407,
|
| 301 |
+
"grad_norm": 644.0,
|
| 302 |
+
"learning_rate": 4.413127538374411e-05,
|
| 303 |
+
"log_odds_chosen": 2.9538180828094482,
|
| 304 |
+
"log_odds_ratio": -0.7920938730239868,
|
| 305 |
+
"logps/chosen": -1.8438174724578857,
|
| 306 |
+
"logps/rejected": -4.690665245056152,
|
| 307 |
+
"loss": 34.7808,
|
| 308 |
+
"nll_loss": 1.8597549200057983,
|
| 309 |
+
"rewards/accuracies": 0.690625011920929,
|
| 310 |
+
"rewards/chosen": -0.9219087362289429,
|
| 311 |
+
"rewards/margins": 1.4234240055084229,
|
| 312 |
+
"rewards/rejected": -2.345332622528076,
|
| 313 |
+
"step": 95
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.9481481481481482,
|
| 317 |
+
"grad_norm": 296.0,
|
| 318 |
+
"learning_rate": 4.320901013934887e-05,
|
| 319 |
+
"log_odds_chosen": 3.595071315765381,
|
| 320 |
+
"log_odds_ratio": -0.8463969230651855,
|
| 321 |
+
"logps/chosen": -2.0342302322387695,
|
| 322 |
+
"logps/rejected": -5.500476837158203,
|
| 323 |
+
"loss": 30.0336,
|
| 324 |
+
"nll_loss": 2.020622491836548,
|
| 325 |
+
"rewards/accuracies": 0.699999988079071,
|
| 326 |
+
"rewards/chosen": -1.0171151161193848,
|
| 327 |
+
"rewards/margins": 1.7331234216690063,
|
| 328 |
+
"rewards/rejected": -2.7502384185791016,
|
| 329 |
+
"step": 100
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.9955555555555555,
|
| 333 |
+
"grad_norm": 312.0,
|
| 334 |
+
"learning_rate": 4.223066054130568e-05,
|
| 335 |
+
"log_odds_chosen": 2.3394980430603027,
|
| 336 |
+
"log_odds_ratio": -0.7494302988052368,
|
| 337 |
+
"logps/chosen": -1.8197135925292969,
|
| 338 |
+
"logps/rejected": -4.058799743652344,
|
| 339 |
+
"loss": 35.016,
|
| 340 |
+
"nll_loss": 1.9035847187042236,
|
| 341 |
+
"rewards/accuracies": 0.668749988079071,
|
| 342 |
+
"rewards/chosen": -0.9098567962646484,
|
| 343 |
+
"rewards/margins": 1.1195429563522339,
|
| 344 |
+
"rewards/rejected": -2.029399871826172,
|
| 345 |
+
"step": 105
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0429629629629629,
|
| 349 |
+
"grad_norm": 1912.0,
|
| 350 |
+
"learning_rate": 4.1199239938743797e-05,
|
| 351 |
+
"log_odds_chosen": 4.236447334289551,
|
| 352 |
+
"log_odds_ratio": -0.4881769120693207,
|
| 353 |
+
"logps/chosen": -2.2219159603118896,
|
| 354 |
+
"logps/rejected": -6.284074306488037,
|
| 355 |
+
"loss": 22.0456,
|
| 356 |
+
"nll_loss": 2.1491634845733643,
|
| 357 |
+
"rewards/accuracies": 0.8031250238418579,
|
| 358 |
+
"rewards/chosen": -1.1109579801559448,
|
| 359 |
+
"rewards/margins": 2.031079053878784,
|
| 360 |
+
"rewards/rejected": -3.1420371532440186,
|
| 361 |
+
"step": 110
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.0903703703703704,
|
| 365 |
+
"grad_norm": 382.0,
|
| 366 |
+
"learning_rate": 4.0117925141242174e-05,
|
| 367 |
+
"log_odds_chosen": 4.837921142578125,
|
| 368 |
+
"log_odds_ratio": -0.5910585522651672,
|
| 369 |
+
"logps/chosen": -1.997674584388733,
|
| 370 |
+
"logps/rejected": -6.637750148773193,
|
| 371 |
+
"loss": 16.3477,
|
| 372 |
+
"nll_loss": 1.9437005519866943,
|
| 373 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 374 |
+
"rewards/chosen": -0.9988372921943665,
|
| 375 |
+
"rewards/margins": 2.320038318634033,
|
| 376 |
+
"rewards/rejected": -3.3188750743865967,
|
| 377 |
+
"step": 115
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.1377777777777778,
|
| 381 |
+
"grad_norm": 516.0,
|
| 382 |
+
"learning_rate": 3.899004663415084e-05,
|
| 383 |
+
"log_odds_chosen": 4.582813739776611,
|
| 384 |
+
"log_odds_ratio": -0.4979763627052307,
|
| 385 |
+
"logps/chosen": -1.5545316934585571,
|
| 386 |
+
"logps/rejected": -5.910370826721191,
|
| 387 |
+
"loss": 11.8075,
|
| 388 |
+
"nll_loss": 1.7471332550048828,
|
| 389 |
+
"rewards/accuracies": 0.784375011920929,
|
| 390 |
+
"rewards/chosen": -0.7772658467292786,
|
| 391 |
+
"rewards/margins": 2.177919626235962,
|
| 392 |
+
"rewards/rejected": -2.9551854133605957,
|
| 393 |
+
"step": 120
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.1851851851851851,
|
| 397 |
+
"grad_norm": 466.0,
|
| 398 |
+
"learning_rate": 3.781907832058587e-05,
|
| 399 |
+
"log_odds_chosen": 5.0084943771362305,
|
| 400 |
+
"log_odds_ratio": -0.5646654367446899,
|
| 401 |
+
"logps/chosen": -1.8939094543457031,
|
| 402 |
+
"logps/rejected": -6.705332279205322,
|
| 403 |
+
"loss": 12.4279,
|
| 404 |
+
"nll_loss": 1.95928156375885,
|
| 405 |
+
"rewards/accuracies": 0.778124988079071,
|
| 406 |
+
"rewards/chosen": -0.9469547271728516,
|
| 407 |
+
"rewards/margins": 2.4057114124298096,
|
| 408 |
+
"rewards/rejected": -3.352666139602661,
|
| 409 |
+
"step": 125
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.2325925925925927,
|
| 413 |
+
"grad_norm": 1016.0,
|
| 414 |
+
"learning_rate": 3.660862682169282e-05,
|
| 415 |
+
"log_odds_chosen": 6.262756824493408,
|
| 416 |
+
"log_odds_ratio": -0.5206202268600464,
|
| 417 |
+
"logps/chosen": -1.6548172235488892,
|
| 418 |
+
"logps/rejected": -7.668360710144043,
|
| 419 |
+
"loss": 2.6131,
|
| 420 |
+
"nll_loss": 1.758933663368225,
|
| 421 |
+
"rewards/accuracies": 0.809374988079071,
|
| 422 |
+
"rewards/chosen": -0.8274086117744446,
|
| 423 |
+
"rewards/margins": 3.006772041320801,
|
| 424 |
+
"rewards/rejected": -3.8341803550720215,
|
| 425 |
+
"step": 130
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.28,
|
| 429 |
+
"grad_norm": 506.0,
|
| 430 |
+
"learning_rate": 3.5362420368134356e-05,
|
| 431 |
+
"log_odds_chosen": 4.817102909088135,
|
| 432 |
+
"log_odds_ratio": -0.6338304281234741,
|
| 433 |
+
"logps/chosen": -1.9214130640029907,
|
| 434 |
+
"logps/rejected": -6.519143104553223,
|
| 435 |
+
"loss": 18.6154,
|
| 436 |
+
"nll_loss": 1.8606481552124023,
|
| 437 |
+
"rewards/accuracies": 0.800000011920929,
|
| 438 |
+
"rewards/chosen": -0.9607065320014954,
|
| 439 |
+
"rewards/margins": 2.2988650798797607,
|
| 440 |
+
"rewards/rejected": -3.2595715522766113,
|
| 441 |
+
"step": 135
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.3274074074074074,
|
| 445 |
+
"grad_norm": 956.0,
|
| 446 |
+
"learning_rate": 3.408429731701635e-05,
|
| 447 |
+
"log_odds_chosen": 4.097002983093262,
|
| 448 |
+
"log_odds_ratio": -0.4646216034889221,
|
| 449 |
+
"logps/chosen": -1.740180253982544,
|
| 450 |
+
"logps/rejected": -5.640007972717285,
|
| 451 |
+
"loss": 11.3345,
|
| 452 |
+
"nll_loss": 1.65927255153656,
|
| 453 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 454 |
+
"rewards/chosen": -0.870090126991272,
|
| 455 |
+
"rewards/margins": 1.9499139785766602,
|
| 456 |
+
"rewards/rejected": -2.8200039863586426,
|
| 457 |
+
"step": 140
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.374814814814815,
|
| 461 |
+
"grad_norm": 1064.0,
|
| 462 |
+
"learning_rate": 3.2778194329621104e-05,
|
| 463 |
+
"log_odds_chosen": 7.051480770111084,
|
| 464 |
+
"log_odds_ratio": -0.4304323196411133,
|
| 465 |
+
"logps/chosen": -2.5247302055358887,
|
| 466 |
+
"logps/rejected": -9.359436988830566,
|
| 467 |
+
"loss": 5.1241,
|
| 468 |
+
"nll_loss": 2.1758809089660645,
|
| 469 |
+
"rewards/accuracies": 0.846875011920929,
|
| 470 |
+
"rewards/chosen": -1.2623651027679443,
|
| 471 |
+
"rewards/margins": 3.417353868484497,
|
| 472 |
+
"rewards/rejected": -4.679718494415283,
|
| 473 |
+
"step": 145
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.4222222222222223,
|
| 477 |
+
"grad_norm": 604.0,
|
| 478 |
+
"learning_rate": 3.144813424636031e-05,
|
| 479 |
+
"log_odds_chosen": 6.2552809715271,
|
| 480 |
+
"log_odds_ratio": -0.4223237931728363,
|
| 481 |
+
"logps/chosen": -1.5957365036010742,
|
| 482 |
+
"logps/rejected": -7.607705116271973,
|
| 483 |
+
"loss": 2.8133,
|
| 484 |
+
"nll_loss": 1.7030588388442993,
|
| 485 |
+
"rewards/accuracies": 0.8125,
|
| 486 |
+
"rewards/chosen": -0.7978682518005371,
|
| 487 |
+
"rewards/margins": 3.005984306335449,
|
| 488 |
+
"rewards/rejected": -3.8038525581359863,
|
| 489 |
+
"step": 150
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.4696296296296296,
|
| 493 |
+
"grad_norm": 972.0,
|
| 494 |
+
"learning_rate": 3.0098213696293542e-05,
|
| 495 |
+
"log_odds_chosen": 5.868264198303223,
|
| 496 |
+
"log_odds_ratio": -0.7328829169273376,
|
| 497 |
+
"logps/chosen": -2.398165225982666,
|
| 498 |
+
"logps/rejected": -8.077600479125977,
|
| 499 |
+
"loss": 14.0782,
|
| 500 |
+
"nll_loss": 1.9348533153533936,
|
| 501 |
+
"rewards/accuracies": 0.78125,
|
| 502 |
+
"rewards/chosen": -1.199082612991333,
|
| 503 |
+
"rewards/margins": 2.839717388153076,
|
| 504 |
+
"rewards/rejected": -4.038800239562988,
|
| 505 |
+
"step": 155
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.5170370370370372,
|
| 509 |
+
"grad_norm": 884.0,
|
| 510 |
+
"learning_rate": 2.8732590479375165e-05,
|
| 511 |
+
"log_odds_chosen": 7.269213676452637,
|
| 512 |
+
"log_odds_ratio": -0.5041804909706116,
|
| 513 |
+
"logps/chosen": -2.1709227561950684,
|
| 514 |
+
"logps/rejected": -9.193258285522461,
|
| 515 |
+
"loss": 1.1204,
|
| 516 |
+
"nll_loss": 2.00057315826416,
|
| 517 |
+
"rewards/accuracies": 0.824999988079071,
|
| 518 |
+
"rewards/chosen": -1.0854613780975342,
|
| 519 |
+
"rewards/margins": 3.511167526245117,
|
| 520 |
+
"rewards/rejected": -4.5966291427612305,
|
| 521 |
+
"step": 160
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.5644444444444443,
|
| 525 |
+
"grad_norm": 580.0,
|
| 526 |
+
"learning_rate": 2.7355470760292956e-05,
|
| 527 |
+
"log_odds_chosen": 6.614538669586182,
|
| 528 |
+
"log_odds_ratio": -0.462197482585907,
|
| 529 |
+
"logps/chosen": -1.875078797340393,
|
| 530 |
+
"logps/rejected": -8.235835075378418,
|
| 531 |
+
"loss": -2.3102,
|
| 532 |
+
"nll_loss": 1.8360283374786377,
|
| 533 |
+
"rewards/accuracies": 0.824999988079071,
|
| 534 |
+
"rewards/chosen": -0.9375393986701965,
|
| 535 |
+
"rewards/margins": 3.1803784370422363,
|
| 536 |
+
"rewards/rejected": -4.117917537689209,
|
| 537 |
+
"step": 165
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.6118518518518519,
|
| 541 |
+
"grad_norm": 1064.0,
|
| 542 |
+
"learning_rate": 2.597109611334169e-05,
|
| 543 |
+
"log_odds_chosen": 6.787781715393066,
|
| 544 |
+
"log_odds_ratio": -0.5965229272842407,
|
| 545 |
+
"logps/chosen": -2.1229805946350098,
|
| 546 |
+
"logps/rejected": -8.671282768249512,
|
| 547 |
+
"loss": -2.3052,
|
| 548 |
+
"nll_loss": 1.9179661273956299,
|
| 549 |
+
"rewards/accuracies": 0.828125,
|
| 550 |
+
"rewards/chosen": -1.0614902973175049,
|
| 551 |
+
"rewards/margins": 3.27415132522583,
|
| 552 |
+
"rewards/rejected": -4.335641384124756,
|
| 553 |
+
"step": 170
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.6592592592592592,
|
| 557 |
+
"grad_norm": 872.0,
|
| 558 |
+
"learning_rate": 2.458373045823404e-05,
|
| 559 |
+
"log_odds_chosen": 6.521327972412109,
|
| 560 |
+
"log_odds_ratio": -0.528164267539978,
|
| 561 |
+
"logps/chosen": -2.1449532508850098,
|
| 562 |
+
"logps/rejected": -8.415470123291016,
|
| 563 |
+
"loss": 1.3161,
|
| 564 |
+
"nll_loss": 2.0013487339019775,
|
| 565 |
+
"rewards/accuracies": 0.809374988079071,
|
| 566 |
+
"rewards/chosen": -1.0724766254425049,
|
| 567 |
+
"rewards/margins": 3.135258197784424,
|
| 568 |
+
"rewards/rejected": -4.207735061645508,
|
| 569 |
+
"step": 175
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7066666666666666,
|
| 573 |
+
"grad_norm": 1256.0,
|
| 574 |
+
"learning_rate": 2.3197646927086697e-05,
|
| 575 |
+
"log_odds_chosen": 7.416574001312256,
|
| 576 |
+
"log_odds_ratio": -0.5437217354774475,
|
| 577 |
+
"logps/chosen": -2.012195587158203,
|
| 578 |
+
"logps/rejected": -9.163311004638672,
|
| 579 |
+
"loss": -6.5787,
|
| 580 |
+
"nll_loss": 1.931572675704956,
|
| 581 |
+
"rewards/accuracies": 0.815625011920929,
|
| 582 |
+
"rewards/chosen": -1.0060977935791016,
|
| 583 |
+
"rewards/margins": 3.575557231903076,
|
| 584 |
+
"rewards/rejected": -4.581655502319336,
|
| 585 |
+
"step": 180
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.7540740740740741,
|
| 589 |
+
"grad_norm": 500.0,
|
| 590 |
+
"learning_rate": 2.1817114703032176e-05,
|
| 591 |
+
"log_odds_chosen": 7.6875762939453125,
|
| 592 |
+
"log_odds_ratio": -0.6660786867141724,
|
| 593 |
+
"logps/chosen": -2.3043220043182373,
|
| 594 |
+
"logps/rejected": -9.788423538208008,
|
| 595 |
+
"loss": -4.8462,
|
| 596 |
+
"nll_loss": 1.9335501194000244,
|
| 597 |
+
"rewards/accuracies": 0.784375011920929,
|
| 598 |
+
"rewards/chosen": -1.1521610021591187,
|
| 599 |
+
"rewards/margins": 3.742050886154175,
|
| 600 |
+
"rewards/rejected": -4.894211769104004,
|
| 601 |
+
"step": 185
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.8014814814814815,
|
| 605 |
+
"grad_norm": 876.0,
|
| 606 |
+
"learning_rate": 2.0446385870993467e-05,
|
| 607 |
+
"log_odds_chosen": 6.940661430358887,
|
| 608 |
+
"log_odds_ratio": -0.4403323233127594,
|
| 609 |
+
"logps/chosen": -1.603877067565918,
|
| 610 |
+
"logps/rejected": -8.270362854003906,
|
| 611 |
+
"loss": -8.6702,
|
| 612 |
+
"nll_loss": 1.6499793529510498,
|
| 613 |
+
"rewards/accuracies": 0.8343750238418579,
|
| 614 |
+
"rewards/chosen": -0.801938533782959,
|
| 615 |
+
"rewards/margins": 3.333242893218994,
|
| 616 |
+
"rewards/rejected": -4.135181427001953,
|
| 617 |
+
"step": 190
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.8488888888888888,
|
| 621 |
+
"grad_norm": 1320.0,
|
| 622 |
+
"learning_rate": 1.9089682321121834e-05,
|
| 623 |
+
"log_odds_chosen": 9.642197608947754,
|
| 624 |
+
"log_odds_ratio": -0.6060672402381897,
|
| 625 |
+
"logps/chosen": -2.7960872650146484,
|
| 626 |
+
"logps/rejected": -12.217751502990723,
|
| 627 |
+
"loss": -4.8162,
|
| 628 |
+
"nll_loss": 2.395156145095825,
|
| 629 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 630 |
+
"rewards/chosen": -1.3980436325073242,
|
| 631 |
+
"rewards/margins": 4.710832595825195,
|
| 632 |
+
"rewards/rejected": -6.108875751495361,
|
| 633 |
+
"step": 195
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.8962962962962964,
|
| 637 |
+
"grad_norm": 1240.0,
|
| 638 |
+
"learning_rate": 1.775118274523545e-05,
|
| 639 |
+
"log_odds_chosen": 6.639898777008057,
|
| 640 |
+
"log_odds_ratio": -0.8482674360275269,
|
| 641 |
+
"logps/chosen": -2.435774564743042,
|
| 642 |
+
"logps/rejected": -8.842093467712402,
|
| 643 |
+
"loss": 3.4566,
|
| 644 |
+
"nll_loss": 1.9974777698516846,
|
| 645 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 646 |
+
"rewards/chosen": -1.217887282371521,
|
| 647 |
+
"rewards/margins": 3.203160047531128,
|
| 648 |
+
"rewards/rejected": -4.421046733856201,
|
| 649 |
+
"step": 200
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.9437037037037037,
|
| 653 |
+
"grad_norm": 900.0,
|
| 654 |
+
"learning_rate": 1.643500976631037e-05,
|
| 655 |
+
"log_odds_chosen": 6.316862106323242,
|
| 656 |
+
"log_odds_ratio": -0.5562707781791687,
|
| 657 |
+
"logps/chosen": -1.6930224895477295,
|
| 658 |
+
"logps/rejected": -7.736858367919922,
|
| 659 |
+
"loss": -2.5638,
|
| 660 |
+
"nll_loss": 1.7886855602264404,
|
| 661 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 662 |
+
"rewards/chosen": -0.8465112447738647,
|
| 663 |
+
"rewards/margins": 3.0219178199768066,
|
| 664 |
+
"rewards/rejected": -3.868429183959961,
|
| 665 |
+
"step": 205
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.991111111111111,
|
| 669 |
+
"grad_norm": 940.0,
|
| 670 |
+
"learning_rate": 1.514521724066537e-05,
|
| 671 |
+
"log_odds_chosen": 6.6161909103393555,
|
| 672 |
+
"log_odds_ratio": -0.8282305598258972,
|
| 673 |
+
"logps/chosen": -2.5031771659851074,
|
| 674 |
+
"logps/rejected": -8.889016151428223,
|
| 675 |
+
"loss": 3.2385,
|
| 676 |
+
"nll_loss": 1.9650413990020752,
|
| 677 |
+
"rewards/accuracies": 0.7749999761581421,
|
| 678 |
+
"rewards/chosen": -1.2515885829925537,
|
| 679 |
+
"rewards/margins": 3.1929194927215576,
|
| 680 |
+
"rewards/rejected": -4.444508075714111,
|
| 681 |
+
"step": 210
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.0385185185185186,
|
| 685 |
+
"grad_norm": 652.0,
|
| 686 |
+
"learning_rate": 1.3885777771950348e-05,
|
| 687 |
+
"log_odds_chosen": 9.489248275756836,
|
| 688 |
+
"log_odds_ratio": -0.48209771513938904,
|
| 689 |
+
"logps/chosen": -1.8140331506729126,
|
| 690 |
+
"logps/rejected": -10.948740005493164,
|
| 691 |
+
"loss": -27.9736,
|
| 692 |
+
"nll_loss": 1.7284278869628906,
|
| 693 |
+
"rewards/accuracies": 0.8374999761581421,
|
| 694 |
+
"rewards/chosen": -0.9070165753364563,
|
| 695 |
+
"rewards/margins": 4.567353248596191,
|
| 696 |
+
"rewards/rejected": -5.474370002746582,
|
| 697 |
+
"step": 215
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.0859259259259257,
|
| 701 |
+
"grad_norm": 1088.0,
|
| 702 |
+
"learning_rate": 1.2660570475395683e-05,
|
| 703 |
+
"log_odds_chosen": 9.268631935119629,
|
| 704 |
+
"log_odds_ratio": -0.5644906759262085,
|
| 705 |
+
"logps/chosen": -2.2069149017333984,
|
| 706 |
+
"logps/rejected": -11.19309139251709,
|
| 707 |
+
"loss": -25.8251,
|
| 708 |
+
"nll_loss": 1.8661258220672607,
|
| 709 |
+
"rewards/accuracies": 0.840624988079071,
|
| 710 |
+
"rewards/chosen": -1.1034574508666992,
|
| 711 |
+
"rewards/margins": 4.493088722229004,
|
| 712 |
+
"rewards/rejected": -5.596545696258545,
|
| 713 |
+
"step": 220
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.1333333333333333,
|
| 717 |
+
"grad_norm": 1536.0,
|
| 718 |
+
"learning_rate": 1.1473369030008974e-05,
|
| 719 |
+
"log_odds_chosen": 13.42963695526123,
|
| 720 |
+
"log_odds_ratio": -0.2936347424983978,
|
| 721 |
+
"logps/chosen": -2.7818703651428223,
|
| 722 |
+
"logps/rejected": -15.93250560760498,
|
| 723 |
+
"loss": -38.3857,
|
| 724 |
+
"nll_loss": 2.1401987075805664,
|
| 725 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 726 |
+
"rewards/chosen": -1.3909351825714111,
|
| 727 |
+
"rewards/margins": 6.5753173828125,
|
| 728 |
+
"rewards/rejected": -7.96625280380249,
|
| 729 |
+
"step": 225
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.180740740740741,
|
| 733 |
+
"grad_norm": 1392.0,
|
| 734 |
+
"learning_rate": 1.0327830055518842e-05,
|
| 735 |
+
"log_odds_chosen": 10.711917877197266,
|
| 736 |
+
"log_odds_ratio": -0.31785595417022705,
|
| 737 |
+
"logps/chosen": -1.934106469154358,
|
| 738 |
+
"logps/rejected": -12.327714920043945,
|
| 739 |
+
"loss": -36.8357,
|
| 740 |
+
"nll_loss": 1.7895313501358032,
|
| 741 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 742 |
+
"rewards/chosen": -0.967053234577179,
|
| 743 |
+
"rewards/margins": 5.196803092956543,
|
| 744 |
+
"rewards/rejected": -6.163857460021973,
|
| 745 |
+
"step": 230
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.228148148148148,
|
| 749 |
+
"grad_norm": 668.0,
|
| 750 |
+
"learning_rate": 9.227481849865235e-06,
|
| 751 |
+
"log_odds_chosen": 10.63416862487793,
|
| 752 |
+
"log_odds_ratio": -0.34915637969970703,
|
| 753 |
+
"logps/chosen": -1.4962049722671509,
|
| 754 |
+
"logps/rejected": -11.764998435974121,
|
| 755 |
+
"loss": -36.5253,
|
| 756 |
+
"nll_loss": 1.6011587381362915,
|
| 757 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 758 |
+
"rewards/chosen": -0.7481024861335754,
|
| 759 |
+
"rewards/margins": 5.134397029876709,
|
| 760 |
+
"rewards/rejected": -5.8824992179870605,
|
| 761 |
+
"step": 235
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 2.2755555555555556,
|
| 765 |
+
"grad_norm": 644.0,
|
| 766 |
+
"learning_rate": 8.175713521924978e-06,
|
| 767 |
+
"log_odds_chosen": 12.23620891571045,
|
| 768 |
+
"log_odds_ratio": -0.1865963637828827,
|
| 769 |
+
"logps/chosen": -1.4724217653274536,
|
| 770 |
+
"logps/rejected": -13.347505569458008,
|
| 771 |
+
"loss": -51.2,
|
| 772 |
+
"nll_loss": 1.510587215423584,
|
| 773 |
+
"rewards/accuracies": 0.8968750238418579,
|
| 774 |
+
"rewards/chosen": -0.7362108826637268,
|
| 775 |
+
"rewards/margins": 5.93754243850708,
|
| 776 |
+
"rewards/rejected": -6.673752784729004,
|
| 777 |
+
"step": 240
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.322962962962963,
|
| 781 |
+
"grad_norm": 1704.0,
|
| 782 |
+
"learning_rate": 7.1757645529443665e-06,
|
| 783 |
+
"log_odds_chosen": 15.528274536132812,
|
| 784 |
+
"log_odds_ratio": -0.33081239461898804,
|
| 785 |
+
"logps/chosen": -2.277860641479492,
|
| 786 |
+
"logps/rejected": -17.496549606323242,
|
| 787 |
+
"loss": -45.7833,
|
| 788 |
+
"nll_loss": 2.0120162963867188,
|
| 789 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 790 |
+
"rewards/chosen": -1.138930320739746,
|
| 791 |
+
"rewards/margins": 7.609343528747559,
|
| 792 |
+
"rewards/rejected": -8.748274803161621,
|
| 793 |
+
"step": 245
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 2.3703703703703702,
|
| 797 |
+
"grad_norm": 1384.0,
|
| 798 |
+
"learning_rate": 6.230714818829733e-06,
|
| 799 |
+
"log_odds_chosen": 14.397786140441895,
|
| 800 |
+
"log_odds_ratio": -0.28585368394851685,
|
| 801 |
+
"logps/chosen": -2.087308645248413,
|
| 802 |
+
"logps/rejected": -16.145244598388672,
|
| 803 |
+
"loss": -41.6852,
|
| 804 |
+
"nll_loss": 1.8240935802459717,
|
| 805 |
+
"rewards/accuracies": 0.90625,
|
| 806 |
+
"rewards/chosen": -1.0436543226242065,
|
| 807 |
+
"rewards/margins": 7.02896785736084,
|
| 808 |
+
"rewards/rejected": -8.072622299194336,
|
| 809 |
+
"step": 250
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.417777777777778,
|
| 813 |
+
"grad_norm": 960.0,
|
| 814 |
+
"learning_rate": 5.343475104027743e-06,
|
| 815 |
+
"log_odds_chosen": 13.305925369262695,
|
| 816 |
+
"log_odds_ratio": -0.2490675449371338,
|
| 817 |
+
"logps/chosen": -1.9038841724395752,
|
| 818 |
+
"logps/rejected": -14.872416496276855,
|
| 819 |
+
"loss": -47.3843,
|
| 820 |
+
"nll_loss": 1.7593896389007568,
|
| 821 |
+
"rewards/accuracies": 0.90625,
|
| 822 |
+
"rewards/chosen": -0.9519420862197876,
|
| 823 |
+
"rewards/margins": 6.484266757965088,
|
| 824 |
+
"rewards/rejected": -7.436208248138428,
|
| 825 |
+
"step": 255
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.4651851851851854,
|
| 829 |
+
"grad_norm": 1152.0,
|
| 830 |
+
"learning_rate": 4.516778136213037e-06,
|
| 831 |
+
"log_odds_chosen": 13.447871208190918,
|
| 832 |
+
"log_odds_ratio": -0.271032452583313,
|
| 833 |
+
"logps/chosen": -1.6334911584854126,
|
| 834 |
+
"logps/rejected": -14.739973068237305,
|
| 835 |
+
"loss": -45.342,
|
| 836 |
+
"nll_loss": 1.6864122152328491,
|
| 837 |
+
"rewards/accuracies": 0.893750011920929,
|
| 838 |
+
"rewards/chosen": -0.8167455792427063,
|
| 839 |
+
"rewards/margins": 6.553240776062012,
|
| 840 |
+
"rewards/rejected": -7.369986534118652,
|
| 841 |
+
"step": 260
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 2.5125925925925925,
|
| 845 |
+
"grad_norm": 1080.0,
|
| 846 |
+
"learning_rate": 3.7531701693965554e-06,
|
| 847 |
+
"log_odds_chosen": 12.9279203414917,
|
| 848 |
+
"log_odds_ratio": -0.3936213552951813,
|
| 849 |
+
"logps/chosen": -1.8595199584960938,
|
| 850 |
+
"logps/rejected": -14.45753288269043,
|
| 851 |
+
"loss": -43.4362,
|
| 852 |
+
"nll_loss": 1.7575229406356812,
|
| 853 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 854 |
+
"rewards/chosen": -0.9297599792480469,
|
| 855 |
+
"rewards/margins": 6.299006938934326,
|
| 856 |
+
"rewards/rejected": -7.228766441345215,
|
| 857 |
+
"step": 265
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.56,
|
| 861 |
+
"grad_norm": 1016.0,
|
| 862 |
+
"learning_rate": 3.055003141378948e-06,
|
| 863 |
+
"log_odds_chosen": 13.242762565612793,
|
| 864 |
+
"log_odds_ratio": -0.2418208122253418,
|
| 865 |
+
"logps/chosen": -1.5649698972702026,
|
| 866 |
+
"logps/rejected": -14.447901725769043,
|
| 867 |
+
"loss": -49.7394,
|
| 868 |
+
"nll_loss": 1.635581612586975,
|
| 869 |
+
"rewards/accuracies": 0.903124988079071,
|
| 870 |
+
"rewards/chosen": -0.7824849486351013,
|
| 871 |
+
"rewards/margins": 6.441465854644775,
|
| 872 |
+
"rewards/rejected": -7.2239508628845215,
|
| 873 |
+
"step": 270
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.6074074074074076,
|
| 877 |
+
"grad_norm": 1080.0,
|
| 878 |
+
"learning_rate": 2.424427429704365e-06,
|
| 879 |
+
"log_odds_chosen": 13.131576538085938,
|
| 880 |
+
"log_odds_ratio": -0.24517115950584412,
|
| 881 |
+
"logps/chosen": -1.7483772039413452,
|
| 882 |
+
"logps/rejected": -14.538126945495605,
|
| 883 |
+
"loss": -50.0705,
|
| 884 |
+
"nll_loss": 1.7109159231185913,
|
| 885 |
+
"rewards/accuracies": 0.918749988079071,
|
| 886 |
+
"rewards/chosen": -0.8741886019706726,
|
| 887 |
+
"rewards/margins": 6.394874572753906,
|
| 888 |
+
"rewards/rejected": -7.269063472747803,
|
| 889 |
+
"step": 275
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 2.6548148148148147,
|
| 893 |
+
"grad_norm": 1320.0,
|
| 894 |
+
"learning_rate": 1.8633852284264508e-06,
|
| 895 |
+
"log_odds_chosen": 12.429898262023926,
|
| 896 |
+
"log_odds_ratio": -0.4006059169769287,
|
| 897 |
+
"logps/chosen": -2.0325827598571777,
|
| 898 |
+
"logps/rejected": -14.141799926757812,
|
| 899 |
+
"loss": -34.9616,
|
| 900 |
+
"nll_loss": 1.9175059795379639,
|
| 901 |
+
"rewards/accuracies": 0.8656250238418579,
|
| 902 |
+
"rewards/chosen": -1.0162913799285889,
|
| 903 |
+
"rewards/margins": 6.054608345031738,
|
| 904 |
+
"rewards/rejected": -7.070899963378906,
|
| 905 |
+
"step": 280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.7022222222222223,
|
| 909 |
+
"grad_norm": 1280.0,
|
| 910 |
+
"learning_rate": 1.3736045660864034e-06,
|
| 911 |
+
"log_odds_chosen": 11.897767066955566,
|
| 912 |
+
"log_odds_ratio": -0.36061421036720276,
|
| 913 |
+
"logps/chosen": -1.7342910766601562,
|
| 914 |
+
"logps/rejected": -13.319345474243164,
|
| 915 |
+
"loss": -42.368,
|
| 916 |
+
"nll_loss": 1.7154544591903687,
|
| 917 |
+
"rewards/accuracies": 0.871874988079071,
|
| 918 |
+
"rewards/chosen": -0.8671455383300781,
|
| 919 |
+
"rewards/margins": 5.792527198791504,
|
| 920 |
+
"rewards/rejected": -6.659672737121582,
|
| 921 |
+
"step": 285
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.74962962962963,
|
| 925 |
+
"grad_norm": 1472.0,
|
| 926 |
+
"learning_rate": 9.565939833279192e-07,
|
| 927 |
+
"log_odds_chosen": 12.41030216217041,
|
| 928 |
+
"log_odds_ratio": -0.48926201462745667,
|
| 929 |
+
"logps/chosen": -2.32023286819458,
|
| 930 |
+
"logps/rejected": -14.431295394897461,
|
| 931 |
+
"loss": -30.8903,
|
| 932 |
+
"nll_loss": 2.0258736610412598,
|
| 933 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 934 |
+
"rewards/chosen": -1.16011643409729,
|
| 935 |
+
"rewards/margins": 6.0555315017700195,
|
| 936 |
+
"rewards/rejected": -7.2156476974487305,
|
| 937 |
+
"step": 290
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 2.797037037037037,
|
| 941 |
+
"grad_norm": 920.0,
|
| 942 |
+
"learning_rate": 6.136378865420872e-07,
|
| 943 |
+
"log_odds_chosen": 12.211552619934082,
|
| 944 |
+
"log_odds_ratio": -0.4335884153842926,
|
| 945 |
+
"logps/chosen": -1.8138420581817627,
|
| 946 |
+
"logps/rejected": -13.713777542114258,
|
| 947 |
+
"loss": -37.5638,
|
| 948 |
+
"nll_loss": 1.7503633499145508,
|
| 949 |
+
"rewards/accuracies": 0.859375,
|
| 950 |
+
"rewards/chosen": -0.9069210290908813,
|
| 951 |
+
"rewards/margins": 5.949967384338379,
|
| 952 |
+
"rewards/rejected": -6.856888771057129,
|
| 953 |
+
"step": 295
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 2.8444444444444446,
|
| 957 |
+
"grad_norm": 1648.0,
|
| 958 |
+
"learning_rate": 3.45792591853214e-07,
|
| 959 |
+
"log_odds_chosen": 13.25987434387207,
|
| 960 |
+
"log_odds_ratio": -0.322501003742218,
|
| 961 |
+
"logps/chosen": -1.6583086252212524,
|
| 962 |
+
"logps/rejected": -14.573458671569824,
|
| 963 |
+
"loss": -45.0139,
|
| 964 |
+
"nll_loss": 1.6782853603363037,
|
| 965 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 966 |
+
"rewards/chosen": -0.8291543126106262,
|
| 967 |
+
"rewards/margins": 6.45757532119751,
|
| 968 |
+
"rewards/rejected": -7.286729335784912,
|
| 969 |
+
"step": 300
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.891851851851852,
|
| 973 |
+
"grad_norm": 1232.0,
|
| 974 |
+
"learning_rate": 1.538830716302092e-07,
|
| 975 |
+
"log_odds_chosen": 12.726137161254883,
|
| 976 |
+
"log_odds_ratio": -0.36502203345298767,
|
| 977 |
+
"logps/chosen": -1.7828433513641357,
|
| 978 |
+
"logps/rejected": -14.167094230651855,
|
| 979 |
+
"loss": -41.4455,
|
| 980 |
+
"nll_loss": 1.751901626586914,
|
| 981 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 982 |
+
"rewards/chosen": -0.8914216756820679,
|
| 983 |
+
"rewards/margins": 6.19212532043457,
|
| 984 |
+
"rewards/rejected": -7.083547115325928,
|
| 985 |
+
"step": 305
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 2.9392592592592592,
|
| 989 |
+
"grad_norm": 1456.0,
|
| 990 |
+
"learning_rate": 3.8500413544415025e-08,
|
| 991 |
+
"log_odds_chosen": 13.806699752807617,
|
| 992 |
+
"log_odds_ratio": -0.15578700602054596,
|
| 993 |
+
"logps/chosen": -1.7442131042480469,
|
| 994 |
+
"logps/rejected": -15.208877563476562,
|
| 995 |
+
"loss": -47.4243,
|
| 996 |
+
"nll_loss": 1.7841014862060547,
|
| 997 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 998 |
+
"rewards/chosen": -0.8721065521240234,
|
| 999 |
+
"rewards/margins": 6.732332706451416,
|
| 1000 |
+
"rewards/rejected": -7.604438781738281,
|
| 1001 |
+
"step": 310
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 2.986666666666667,
|
| 1005 |
+
"grad_norm": 1952.0,
|
| 1006 |
+
"learning_rate": 0.0,
|
| 1007 |
+
"log_odds_chosen": 12.691495895385742,
|
| 1008 |
+
"log_odds_ratio": -0.2689394950866699,
|
| 1009 |
+
"logps/chosen": -1.6713707447052002,
|
| 1010 |
+
"logps/rejected": -13.993799209594727,
|
| 1011 |
+
"loss": -43.2258,
|
| 1012 |
+
"nll_loss": 1.702843427658081,
|
| 1013 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 1014 |
+
"rewards/chosen": -0.8356853723526001,
|
| 1015 |
+
"rewards/margins": 6.1612138748168945,
|
| 1016 |
+
"rewards/rejected": -6.996899604797363,
|
| 1017 |
+
"step": 315
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.986666666666667,
|
| 1021 |
+
"step": 315,
|
| 1022 |
+
"total_flos": 0.0,
|
| 1023 |
+
"train_loss": 25.709435809604706,
|
| 1024 |
+
"train_runtime": 7268.7111,
|
| 1025 |
+
"train_samples_per_second": 2.786,
|
| 1026 |
+
"train_steps_per_second": 0.043
|
| 1027 |
+
}
|
| 1028 |
+
],
|
| 1029 |
+
"logging_steps": 5,
|
| 1030 |
+
"max_steps": 315,
|
| 1031 |
+
"num_input_tokens_seen": 0,
|
| 1032 |
+
"num_train_epochs": 3,
|
| 1033 |
+
"save_steps": 100000,
|
| 1034 |
+
"stateful_callbacks": {
|
| 1035 |
+
"TrainerControl": {
|
| 1036 |
+
"args": {
|
| 1037 |
+
"should_epoch_stop": false,
|
| 1038 |
+
"should_evaluate": false,
|
| 1039 |
+
"should_log": false,
|
| 1040 |
+
"should_save": true,
|
| 1041 |
+
"should_training_stop": true
|
| 1042 |
+
},
|
| 1043 |
+
"attributes": {}
|
| 1044 |
+
}
|
| 1045 |
+
},
|
| 1046 |
+
"total_flos": 0.0,
|
| 1047 |
+
"train_batch_size": 1,
|
| 1048 |
+
"trial_name": null,
|
| 1049 |
+
"trial_params": null
|
| 1050 |
+
}
|