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-v5-1sftt-gamma10
|
| 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-v5-1sftt-gamma10
|
| 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-v5-1sftt-gamma10", 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/hmm7jfjo)
|
| 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": 44.452726506430004,
|
| 5 |
+
"train_runtime": 7306.7465,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.771,
|
| 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": 44.452726506430004,
|
| 5 |
+
"train_runtime": 7306.7465,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.771,
|
| 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": 1128.0,
|
| 14 |
+
"learning_rate": 7.8125e-06,
|
| 15 |
+
"log_odds_chosen": -1.058196783065796,
|
| 16 |
+
"log_odds_ratio": -11.008082389831543,
|
| 17 |
+
"logps/chosen": -22.712032318115234,
|
| 18 |
+
"logps/rejected": -21.654064178466797,
|
| 19 |
+
"loss": 467.0638,
|
| 20 |
+
"nll_loss": 9.837631225585938,
|
| 21 |
+
"rewards/accuracies": 0.4781250059604645,
|
| 22 |
+
"rewards/chosen": -11.356016159057617,
|
| 23 |
+
"rewards/margins": -0.5289839506149292,
|
| 24 |
+
"rewards/rejected": -10.827032089233398,
|
| 25 |
+
"step": 5
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.09481481481481481,
|
| 29 |
+
"grad_norm": 588.0,
|
| 30 |
+
"learning_rate": 1.5625e-05,
|
| 31 |
+
"log_odds_chosen": -3.1185920238494873,
|
| 32 |
+
"log_odds_ratio": -11.042587280273438,
|
| 33 |
+
"logps/chosen": -21.993783950805664,
|
| 34 |
+
"logps/rejected": -18.875520706176758,
|
| 35 |
+
"loss": 435.3677,
|
| 36 |
+
"nll_loss": 8.753497123718262,
|
| 37 |
+
"rewards/accuracies": 0.453125,
|
| 38 |
+
"rewards/chosen": -10.996891975402832,
|
| 39 |
+
"rewards/margins": -1.5591304302215576,
|
| 40 |
+
"rewards/rejected": -9.437760353088379,
|
| 41 |
+
"step": 10
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.14222222222222222,
|
| 45 |
+
"grad_norm": 468.0,
|
| 46 |
+
"learning_rate": 2.34375e-05,
|
| 47 |
+
"log_odds_chosen": -1.2942270040512085,
|
| 48 |
+
"log_odds_ratio": -11.319032669067383,
|
| 49 |
+
"logps/chosen": -21.984771728515625,
|
| 50 |
+
"logps/rejected": -20.689090728759766,
|
| 51 |
+
"loss": 412.2578,
|
| 52 |
+
"nll_loss": 7.999310493469238,
|
| 53 |
+
"rewards/accuracies": 0.515625,
|
| 54 |
+
"rewards/chosen": -10.992385864257812,
|
| 55 |
+
"rewards/margins": -0.6478394269943237,
|
| 56 |
+
"rewards/rejected": -10.344545364379883,
|
| 57 |
+
"step": 15
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.18962962962962962,
|
| 61 |
+
"grad_norm": 8384.0,
|
| 62 |
+
"learning_rate": 3.125e-05,
|
| 63 |
+
"log_odds_chosen": -4.4439616203308105,
|
| 64 |
+
"log_odds_ratio": -9.70390510559082,
|
| 65 |
+
"logps/chosen": -17.518526077270508,
|
| 66 |
+
"logps/rejected": -13.076098442077637,
|
| 67 |
+
"loss": 338.9335,
|
| 68 |
+
"nll_loss": 6.398016929626465,
|
| 69 |
+
"rewards/accuracies": 0.4281249940395355,
|
| 70 |
+
"rewards/chosen": -8.759263038635254,
|
| 71 |
+
"rewards/margins": -2.2212135791778564,
|
| 72 |
+
"rewards/rejected": -6.538049221038818,
|
| 73 |
+
"step": 20
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.23703703703703705,
|
| 77 |
+
"grad_norm": 242.0,
|
| 78 |
+
"learning_rate": 3.90625e-05,
|
| 79 |
+
"log_odds_chosen": -0.14175763726234436,
|
| 80 |
+
"log_odds_ratio": -2.161510944366455,
|
| 81 |
+
"logps/chosen": -4.730892658233643,
|
| 82 |
+
"logps/rejected": -4.572934150695801,
|
| 83 |
+
"loss": 107.5018,
|
| 84 |
+
"nll_loss": 2.8825061321258545,
|
| 85 |
+
"rewards/accuracies": 0.543749988079071,
|
| 86 |
+
"rewards/chosen": -2.3654463291168213,
|
| 87 |
+
"rewards/margins": -0.07897911965847015,
|
| 88 |
+
"rewards/rejected": -2.2864670753479004,
|
| 89 |
+
"step": 25
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.28444444444444444,
|
| 93 |
+
"grad_norm": 186.0,
|
| 94 |
+
"learning_rate": 4.6875e-05,
|
| 95 |
+
"log_odds_chosen": 0.1054491400718689,
|
| 96 |
+
"log_odds_ratio": -0.8509915471076965,
|
| 97 |
+
"logps/chosen": -1.8616880178451538,
|
| 98 |
+
"logps/rejected": -1.9526846408843994,
|
| 99 |
+
"loss": 65.1113,
|
| 100 |
+
"nll_loss": 2.0738651752471924,
|
| 101 |
+
"rewards/accuracies": 0.543749988079071,
|
| 102 |
+
"rewards/chosen": -0.9308440089225769,
|
| 103 |
+
"rewards/margins": 0.04549835994839668,
|
| 104 |
+
"rewards/rejected": -0.9763423204421997,
|
| 105 |
+
"step": 30
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33185185185185184,
|
| 109 |
+
"grad_norm": 205.0,
|
| 110 |
+
"learning_rate": 4.998613757348784e-05,
|
| 111 |
+
"log_odds_chosen": 0.31355661153793335,
|
| 112 |
+
"log_odds_ratio": -0.7803260684013367,
|
| 113 |
+
"logps/chosen": -1.765001654624939,
|
| 114 |
+
"logps/rejected": -2.0435233116149902,
|
| 115 |
+
"loss": 58.3669,
|
| 116 |
+
"nll_loss": 1.9382717609405518,
|
| 117 |
+
"rewards/accuracies": 0.550000011920929,
|
| 118 |
+
"rewards/chosen": -0.8825008273124695,
|
| 119 |
+
"rewards/margins": 0.1392608880996704,
|
| 120 |
+
"rewards/rejected": -1.0217616558074951,
|
| 121 |
+
"step": 35
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.37925925925925924,
|
| 125 |
+
"grad_norm": 167.0,
|
| 126 |
+
"learning_rate": 4.990147841143462e-05,
|
| 127 |
+
"log_odds_chosen": 0.5204086899757385,
|
| 128 |
+
"log_odds_ratio": -0.653488278388977,
|
| 129 |
+
"logps/chosen": -1.6190803050994873,
|
| 130 |
+
"logps/rejected": -2.0669641494750977,
|
| 131 |
+
"loss": 53.3821,
|
| 132 |
+
"nll_loss": 1.8854446411132812,
|
| 133 |
+
"rewards/accuracies": 0.6656249761581421,
|
| 134 |
+
"rewards/chosen": -0.8095401525497437,
|
| 135 |
+
"rewards/margins": 0.22394177317619324,
|
| 136 |
+
"rewards/rejected": -1.0334820747375488,
|
| 137 |
+
"step": 40
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.4266666666666667,
|
| 141 |
+
"grad_norm": 350.0,
|
| 142 |
+
"learning_rate": 4.97401218720448e-05,
|
| 143 |
+
"log_odds_chosen": 1.1803003549575806,
|
| 144 |
+
"log_odds_ratio": -0.5899690389633179,
|
| 145 |
+
"logps/chosen": -1.6516567468643188,
|
| 146 |
+
"logps/rejected": -2.7521023750305176,
|
| 147 |
+
"loss": 46.1493,
|
| 148 |
+
"nll_loss": 1.8530938625335693,
|
| 149 |
+
"rewards/accuracies": 0.6968749761581421,
|
| 150 |
+
"rewards/chosen": -0.8258283734321594,
|
| 151 |
+
"rewards/margins": 0.5502227544784546,
|
| 152 |
+
"rewards/rejected": -1.3760511875152588,
|
| 153 |
+
"step": 45
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.4740740740740741,
|
| 157 |
+
"grad_norm": 230.0,
|
| 158 |
+
"learning_rate": 4.9502564938797946e-05,
|
| 159 |
+
"log_odds_chosen": 0.8049520254135132,
|
| 160 |
+
"log_odds_ratio": -0.6416460871696472,
|
| 161 |
+
"logps/chosen": -1.554616928100586,
|
| 162 |
+
"logps/rejected": -2.283681869506836,
|
| 163 |
+
"loss": 48.6877,
|
| 164 |
+
"nll_loss": 1.7986338138580322,
|
| 165 |
+
"rewards/accuracies": 0.659375011920929,
|
| 166 |
+
"rewards/chosen": -0.777308464050293,
|
| 167 |
+
"rewards/margins": 0.36453238129615784,
|
| 168 |
+
"rewards/rejected": -1.141840934753418,
|
| 169 |
+
"step": 50
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.5214814814814814,
|
| 173 |
+
"grad_norm": 155.0,
|
| 174 |
+
"learning_rate": 4.918953929490768e-05,
|
| 175 |
+
"log_odds_chosen": 1.1016016006469727,
|
| 176 |
+
"log_odds_ratio": -0.6536484360694885,
|
| 177 |
+
"logps/chosen": -1.6193764209747314,
|
| 178 |
+
"logps/rejected": -2.638171434402466,
|
| 179 |
+
"loss": 47.3989,
|
| 180 |
+
"nll_loss": 1.8205287456512451,
|
| 181 |
+
"rewards/accuracies": 0.6656249761581421,
|
| 182 |
+
"rewards/chosen": -0.8096882104873657,
|
| 183 |
+
"rewards/margins": 0.5093975067138672,
|
| 184 |
+
"rewards/rejected": -1.319085717201233,
|
| 185 |
+
"step": 55
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.5688888888888889,
|
| 189 |
+
"grad_norm": 348.0,
|
| 190 |
+
"learning_rate": 4.88020090697132e-05,
|
| 191 |
+
"log_odds_chosen": 1.6798295974731445,
|
| 192 |
+
"log_odds_ratio": -0.6031836271286011,
|
| 193 |
+
"logps/chosen": -1.5959837436676025,
|
| 194 |
+
"logps/rejected": -3.1869189739227295,
|
| 195 |
+
"loss": 46.0557,
|
| 196 |
+
"nll_loss": 1.8288695812225342,
|
| 197 |
+
"rewards/accuracies": 0.6875,
|
| 198 |
+
"rewards/chosen": -0.7979918718338013,
|
| 199 |
+
"rewards/margins": 0.7954676151275635,
|
| 200 |
+
"rewards/rejected": -1.5934594869613647,
|
| 201 |
+
"step": 60
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.6162962962962963,
|
| 205 |
+
"grad_norm": 155.0,
|
| 206 |
+
"learning_rate": 4.834116786912897e-05,
|
| 207 |
+
"log_odds_chosen": 1.7133772373199463,
|
| 208 |
+
"log_odds_ratio": -0.7503857016563416,
|
| 209 |
+
"logps/chosen": -1.8493322134017944,
|
| 210 |
+
"logps/rejected": -3.4778411388397217,
|
| 211 |
+
"loss": 52.1616,
|
| 212 |
+
"nll_loss": 1.9849662780761719,
|
| 213 |
+
"rewards/accuracies": 0.671875,
|
| 214 |
+
"rewards/chosen": -0.9246661067008972,
|
| 215 |
+
"rewards/margins": 0.8142545819282532,
|
| 216 |
+
"rewards/rejected": -1.7389205694198608,
|
| 217 |
+
"step": 65
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.6637037037037037,
|
| 221 |
+
"grad_norm": 348.0,
|
| 222 |
+
"learning_rate": 4.7808435099299045e-05,
|
| 223 |
+
"log_odds_chosen": 1.9798109531402588,
|
| 224 |
+
"log_odds_ratio": -0.5159353017807007,
|
| 225 |
+
"logps/chosen": -1.7272640466690063,
|
| 226 |
+
"logps/rejected": -3.5779659748077393,
|
| 227 |
+
"loss": 38.9021,
|
| 228 |
+
"nll_loss": 1.842188835144043,
|
| 229 |
+
"rewards/accuracies": 0.7718750238418579,
|
| 230 |
+
"rewards/chosen": -0.8636320233345032,
|
| 231 |
+
"rewards/margins": 0.9253507852554321,
|
| 232 |
+
"rewards/rejected": -1.7889829874038696,
|
| 233 |
+
"step": 70
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.7111111111111111,
|
| 237 |
+
"grad_norm": 224.0,
|
| 238 |
+
"learning_rate": 4.720545159477922e-05,
|
| 239 |
+
"log_odds_chosen": 1.3524739742279053,
|
| 240 |
+
"log_odds_ratio": -0.6269196271896362,
|
| 241 |
+
"logps/chosen": -1.5649521350860596,
|
| 242 |
+
"logps/rejected": -2.8103561401367188,
|
| 243 |
+
"loss": 42.7284,
|
| 244 |
+
"nll_loss": 1.7544043064117432,
|
| 245 |
+
"rewards/accuracies": 0.7093750238418579,
|
| 246 |
+
"rewards/chosen": -0.7824760675430298,
|
| 247 |
+
"rewards/margins": 0.6227020025253296,
|
| 248 |
+
"rewards/rejected": -1.4051780700683594,
|
| 249 |
+
"step": 75
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.7585185185185185,
|
| 253 |
+
"grad_norm": 318.0,
|
| 254 |
+
"learning_rate": 4.653407456471222e-05,
|
| 255 |
+
"log_odds_chosen": 1.2641541957855225,
|
| 256 |
+
"log_odds_ratio": -0.7116214036941528,
|
| 257 |
+
"logps/chosen": -1.672009825706482,
|
| 258 |
+
"logps/rejected": -2.8201100826263428,
|
| 259 |
+
"loss": 45.8222,
|
| 260 |
+
"nll_loss": 1.8372602462768555,
|
| 261 |
+
"rewards/accuracies": 0.640625,
|
| 262 |
+
"rewards/chosen": -0.836004912853241,
|
| 263 |
+
"rewards/margins": 0.57405024766922,
|
| 264 |
+
"rewards/rejected": -1.4100550413131714,
|
| 265 |
+
"step": 80
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.8059259259259259,
|
| 269 |
+
"grad_norm": 812.0,
|
| 270 |
+
"learning_rate": 4.579637187256222e-05,
|
| 271 |
+
"log_odds_chosen": 2.119652509689331,
|
| 272 |
+
"log_odds_ratio": -0.7182853817939758,
|
| 273 |
+
"logps/chosen": -1.69955575466156,
|
| 274 |
+
"logps/rejected": -3.706132411956787,
|
| 275 |
+
"loss": 44.067,
|
| 276 |
+
"nll_loss": 1.8409528732299805,
|
| 277 |
+
"rewards/accuracies": 0.675000011920929,
|
| 278 |
+
"rewards/chosen": -0.84977787733078,
|
| 279 |
+
"rewards/margins": 1.0032880306243896,
|
| 280 |
+
"rewards/rejected": -1.8530662059783936,
|
| 281 |
+
"step": 85
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.8533333333333334,
|
| 285 |
+
"grad_norm": 632.0,
|
| 286 |
+
"learning_rate": 4.499461566702685e-05,
|
| 287 |
+
"log_odds_chosen": 3.368741512298584,
|
| 288 |
+
"log_odds_ratio": -0.9751121401786804,
|
| 289 |
+
"logps/chosen": -1.8076425790786743,
|
| 290 |
+
"logps/rejected": -5.029626369476318,
|
| 291 |
+
"loss": 48.4689,
|
| 292 |
+
"nll_loss": 1.954244613647461,
|
| 293 |
+
"rewards/accuracies": 0.715624988079071,
|
| 294 |
+
"rewards/chosen": -0.9038212895393372,
|
| 295 |
+
"rewards/margins": 1.6109920740127563,
|
| 296 |
+
"rewards/rejected": -2.514813184738159,
|
| 297 |
+
"step": 90
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.9007407407407407,
|
| 301 |
+
"grad_norm": 354.0,
|
| 302 |
+
"learning_rate": 4.413127538374411e-05,
|
| 303 |
+
"log_odds_chosen": 0.9373680353164673,
|
| 304 |
+
"log_odds_ratio": -0.6437951922416687,
|
| 305 |
+
"logps/chosen": -1.4754348993301392,
|
| 306 |
+
"logps/rejected": -2.3128228187561035,
|
| 307 |
+
"loss": 42.7029,
|
| 308 |
+
"nll_loss": 1.6802858114242554,
|
| 309 |
+
"rewards/accuracies": 0.6468750238418579,
|
| 310 |
+
"rewards/chosen": -0.7377174496650696,
|
| 311 |
+
"rewards/margins": 0.418693870306015,
|
| 312 |
+
"rewards/rejected": -1.1564114093780518,
|
| 313 |
+
"step": 95
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.9481481481481482,
|
| 317 |
+
"grad_norm": 292.0,
|
| 318 |
+
"learning_rate": 4.320901013934887e-05,
|
| 319 |
+
"log_odds_chosen": 1.350513219833374,
|
| 320 |
+
"log_odds_ratio": -0.5488158464431763,
|
| 321 |
+
"logps/chosen": -1.4655238389968872,
|
| 322 |
+
"logps/rejected": -2.687412738800049,
|
| 323 |
+
"loss": 37.986,
|
| 324 |
+
"nll_loss": 1.686754822731018,
|
| 325 |
+
"rewards/accuracies": 0.703125,
|
| 326 |
+
"rewards/chosen": -0.7327619194984436,
|
| 327 |
+
"rewards/margins": 0.610944390296936,
|
| 328 |
+
"rewards/rejected": -1.3437063694000244,
|
| 329 |
+
"step": 100
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.9955555555555555,
|
| 333 |
+
"grad_norm": 167.0,
|
| 334 |
+
"learning_rate": 4.223066054130568e-05,
|
| 335 |
+
"log_odds_chosen": 1.7379270792007446,
|
| 336 |
+
"log_odds_ratio": -0.6734561324119568,
|
| 337 |
+
"logps/chosen": -2.0734519958496094,
|
| 338 |
+
"logps/rejected": -3.697786808013916,
|
| 339 |
+
"loss": 47.1636,
|
| 340 |
+
"nll_loss": 2.014409065246582,
|
| 341 |
+
"rewards/accuracies": 0.6781250238418579,
|
| 342 |
+
"rewards/chosen": -1.0367259979248047,
|
| 343 |
+
"rewards/margins": 0.8121673464775085,
|
| 344 |
+
"rewards/rejected": -1.848893404006958,
|
| 345 |
+
"step": 105
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0429629629629629,
|
| 349 |
+
"grad_norm": 544.0,
|
| 350 |
+
"learning_rate": 4.1199239938743797e-05,
|
| 351 |
+
"log_odds_chosen": 3.0268468856811523,
|
| 352 |
+
"log_odds_ratio": -0.44861817359924316,
|
| 353 |
+
"logps/chosen": -1.4808493852615356,
|
| 354 |
+
"logps/rejected": -4.2748918533325195,
|
| 355 |
+
"loss": 27.2254,
|
| 356 |
+
"nll_loss": 1.616323709487915,
|
| 357 |
+
"rewards/accuracies": 0.78125,
|
| 358 |
+
"rewards/chosen": -0.7404246926307678,
|
| 359 |
+
"rewards/margins": 1.3970211744308472,
|
| 360 |
+
"rewards/rejected": -2.1374459266662598,
|
| 361 |
+
"step": 110
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.0903703703703704,
|
| 365 |
+
"grad_norm": 193.0,
|
| 366 |
+
"learning_rate": 4.0117925141242174e-05,
|
| 367 |
+
"log_odds_chosen": 2.6167469024658203,
|
| 368 |
+
"log_odds_ratio": -0.4948674738407135,
|
| 369 |
+
"logps/chosen": -1.3744416236877441,
|
| 370 |
+
"logps/rejected": -3.7343673706054688,
|
| 371 |
+
"loss": 25.6633,
|
| 372 |
+
"nll_loss": 1.5660358667373657,
|
| 373 |
+
"rewards/accuracies": 0.800000011920929,
|
| 374 |
+
"rewards/chosen": -0.6872208118438721,
|
| 375 |
+
"rewards/margins": 1.1799629926681519,
|
| 376 |
+
"rewards/rejected": -1.8671836853027344,
|
| 377 |
+
"step": 115
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.1377777777777778,
|
| 381 |
+
"grad_norm": 358.0,
|
| 382 |
+
"learning_rate": 3.899004663415084e-05,
|
| 383 |
+
"log_odds_chosen": 2.8587183952331543,
|
| 384 |
+
"log_odds_ratio": -0.4674352705478668,
|
| 385 |
+
"logps/chosen": -1.3489106893539429,
|
| 386 |
+
"logps/rejected": -3.9463298320770264,
|
| 387 |
+
"loss": 24.1664,
|
| 388 |
+
"nll_loss": 1.5956518650054932,
|
| 389 |
+
"rewards/accuracies": 0.8062499761581421,
|
| 390 |
+
"rewards/chosen": -0.6744553446769714,
|
| 391 |
+
"rewards/margins": 1.2987096309661865,
|
| 392 |
+
"rewards/rejected": -1.9731649160385132,
|
| 393 |
+
"step": 120
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.1851851851851851,
|
| 397 |
+
"grad_norm": 228.0,
|
| 398 |
+
"learning_rate": 3.781907832058587e-05,
|
| 399 |
+
"log_odds_chosen": 2.298691511154175,
|
| 400 |
+
"log_odds_ratio": -0.44407153129577637,
|
| 401 |
+
"logps/chosen": -1.3952347040176392,
|
| 402 |
+
"logps/rejected": -3.443612575531006,
|
| 403 |
+
"loss": 26.9036,
|
| 404 |
+
"nll_loss": 1.621145248413086,
|
| 405 |
+
"rewards/accuracies": 0.793749988079071,
|
| 406 |
+
"rewards/chosen": -0.6976173520088196,
|
| 407 |
+
"rewards/margins": 1.0241888761520386,
|
| 408 |
+
"rewards/rejected": -1.721806287765503,
|
| 409 |
+
"step": 125
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.2325925925925927,
|
| 413 |
+
"grad_norm": 470.0,
|
| 414 |
+
"learning_rate": 3.660862682169282e-05,
|
| 415 |
+
"log_odds_chosen": 3.3581974506378174,
|
| 416 |
+
"log_odds_ratio": -0.37174850702285767,
|
| 417 |
+
"logps/chosen": -1.641959547996521,
|
| 418 |
+
"logps/rejected": -4.736272811889648,
|
| 419 |
+
"loss": 19.7077,
|
| 420 |
+
"nll_loss": 1.707945466041565,
|
| 421 |
+
"rewards/accuracies": 0.828125,
|
| 422 |
+
"rewards/chosen": -0.8209797739982605,
|
| 423 |
+
"rewards/margins": 1.5471566915512085,
|
| 424 |
+
"rewards/rejected": -2.368136405944824,
|
| 425 |
+
"step": 130
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.28,
|
| 429 |
+
"grad_norm": 260.0,
|
| 430 |
+
"learning_rate": 3.5362420368134356e-05,
|
| 431 |
+
"log_odds_chosen": 2.5022172927856445,
|
| 432 |
+
"log_odds_ratio": -0.4535306990146637,
|
| 433 |
+
"logps/chosen": -1.4220046997070312,
|
| 434 |
+
"logps/rejected": -3.6646904945373535,
|
| 435 |
+
"loss": 25.0471,
|
| 436 |
+
"nll_loss": 1.5830011367797852,
|
| 437 |
+
"rewards/accuracies": 0.824999988079071,
|
| 438 |
+
"rewards/chosen": -0.7110023498535156,
|
| 439 |
+
"rewards/margins": 1.1213428974151611,
|
| 440 |
+
"rewards/rejected": -1.8323452472686768,
|
| 441 |
+
"step": 135
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.3274074074074074,
|
| 445 |
+
"grad_norm": 484.0,
|
| 446 |
+
"learning_rate": 3.408429731701635e-05,
|
| 447 |
+
"log_odds_chosen": 2.461456775665283,
|
| 448 |
+
"log_odds_ratio": -0.4378852844238281,
|
| 449 |
+
"logps/chosen": -1.438861608505249,
|
| 450 |
+
"logps/rejected": -3.6586761474609375,
|
| 451 |
+
"loss": 25.1637,
|
| 452 |
+
"nll_loss": 1.5763061046600342,
|
| 453 |
+
"rewards/accuracies": 0.815625011920929,
|
| 454 |
+
"rewards/chosen": -0.7194308042526245,
|
| 455 |
+
"rewards/margins": 1.1099072694778442,
|
| 456 |
+
"rewards/rejected": -1.8293380737304688,
|
| 457 |
+
"step": 140
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.374814814814815,
|
| 461 |
+
"grad_norm": 366.0,
|
| 462 |
+
"learning_rate": 3.2778194329621104e-05,
|
| 463 |
+
"log_odds_chosen": 3.443436861038208,
|
| 464 |
+
"log_odds_ratio": -0.3327510952949524,
|
| 465 |
+
"logps/chosen": -1.639411211013794,
|
| 466 |
+
"logps/rejected": -4.814173221588135,
|
| 467 |
+
"loss": 19.2385,
|
| 468 |
+
"nll_loss": 1.6921018362045288,
|
| 469 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 470 |
+
"rewards/chosen": -0.819705605506897,
|
| 471 |
+
"rewards/margins": 1.5873807668685913,
|
| 472 |
+
"rewards/rejected": -2.4070866107940674,
|
| 473 |
+
"step": 145
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.4222222222222223,
|
| 477 |
+
"grad_norm": 460.0,
|
| 478 |
+
"learning_rate": 3.144813424636031e-05,
|
| 479 |
+
"log_odds_chosen": 3.7978568077087402,
|
| 480 |
+
"log_odds_ratio": -0.3528765141963959,
|
| 481 |
+
"logps/chosen": -1.41143000125885,
|
| 482 |
+
"logps/rejected": -4.935047149658203,
|
| 483 |
+
"loss": 16.6229,
|
| 484 |
+
"nll_loss": 1.5774896144866943,
|
| 485 |
+
"rewards/accuracies": 0.846875011920929,
|
| 486 |
+
"rewards/chosen": -0.705715000629425,
|
| 487 |
+
"rewards/margins": 1.7618083953857422,
|
| 488 |
+
"rewards/rejected": -2.4675235748291016,
|
| 489 |
+
"step": 150
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.4696296296296296,
|
| 493 |
+
"grad_norm": 414.0,
|
| 494 |
+
"learning_rate": 3.0098213696293542e-05,
|
| 495 |
+
"log_odds_chosen": 3.518047332763672,
|
| 496 |
+
"log_odds_ratio": -0.5385881662368774,
|
| 497 |
+
"logps/chosen": -1.7592281103134155,
|
| 498 |
+
"logps/rejected": -5.035131931304932,
|
| 499 |
+
"loss": 23.1146,
|
| 500 |
+
"nll_loss": 1.7055237293243408,
|
| 501 |
+
"rewards/accuracies": 0.8125,
|
| 502 |
+
"rewards/chosen": -0.8796140551567078,
|
| 503 |
+
"rewards/margins": 1.6379520893096924,
|
| 504 |
+
"rewards/rejected": -2.517565965652466,
|
| 505 |
+
"step": 155
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.5170370370370372,
|
| 509 |
+
"grad_norm": 688.0,
|
| 510 |
+
"learning_rate": 2.8732590479375165e-05,
|
| 511 |
+
"log_odds_chosen": 3.530190944671631,
|
| 512 |
+
"log_odds_ratio": -0.4366869032382965,
|
| 513 |
+
"logps/chosen": -1.6791108846664429,
|
| 514 |
+
"logps/rejected": -4.922083854675293,
|
| 515 |
+
"loss": 17.4724,
|
| 516 |
+
"nll_loss": 1.6354753971099854,
|
| 517 |
+
"rewards/accuracies": 0.809374988079071,
|
| 518 |
+
"rewards/chosen": -0.8395554423332214,
|
| 519 |
+
"rewards/margins": 1.6214866638183594,
|
| 520 |
+
"rewards/rejected": -2.4610419273376465,
|
| 521 |
+
"step": 160
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.5644444444444443,
|
| 525 |
+
"grad_norm": 288.0,
|
| 526 |
+
"learning_rate": 2.7355470760292956e-05,
|
| 527 |
+
"log_odds_chosen": 3.844754695892334,
|
| 528 |
+
"log_odds_ratio": -0.3585518002510071,
|
| 529 |
+
"logps/chosen": -1.4751592874526978,
|
| 530 |
+
"logps/rejected": -5.015917778015137,
|
| 531 |
+
"loss": 14.768,
|
| 532 |
+
"nll_loss": 1.5972576141357422,
|
| 533 |
+
"rewards/accuracies": 0.846875011920929,
|
| 534 |
+
"rewards/chosen": -0.7375796437263489,
|
| 535 |
+
"rewards/margins": 1.7703793048858643,
|
| 536 |
+
"rewards/rejected": -2.5079588890075684,
|
| 537 |
+
"step": 165
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.6118518518518519,
|
| 541 |
+
"grad_norm": 388.0,
|
| 542 |
+
"learning_rate": 2.597109611334169e-05,
|
| 543 |
+
"log_odds_chosen": 4.079934597015381,
|
| 544 |
+
"log_odds_ratio": -0.44897276163101196,
|
| 545 |
+
"logps/chosen": -1.6552518606185913,
|
| 546 |
+
"logps/rejected": -5.458821773529053,
|
| 547 |
+
"loss": 15.6283,
|
| 548 |
+
"nll_loss": 1.6442257165908813,
|
| 549 |
+
"rewards/accuracies": 0.815625011920929,
|
| 550 |
+
"rewards/chosen": -0.8276259303092957,
|
| 551 |
+
"rewards/margins": 1.901785135269165,
|
| 552 |
+
"rewards/rejected": -2.7294108867645264,
|
| 553 |
+
"step": 170
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.6592592592592592,
|
| 557 |
+
"grad_norm": 356.0,
|
| 558 |
+
"learning_rate": 2.458373045823404e-05,
|
| 559 |
+
"log_odds_chosen": 3.561838150024414,
|
| 560 |
+
"log_odds_ratio": -0.3351458013057709,
|
| 561 |
+
"logps/chosen": -1.2622171640396118,
|
| 562 |
+
"logps/rejected": -4.530435085296631,
|
| 563 |
+
"loss": 14.9451,
|
| 564 |
+
"nll_loss": 1.494937777519226,
|
| 565 |
+
"rewards/accuracies": 0.828125,
|
| 566 |
+
"rewards/chosen": -0.6311085820198059,
|
| 567 |
+
"rewards/margins": 1.6341091394424438,
|
| 568 |
+
"rewards/rejected": -2.2652175426483154,
|
| 569 |
+
"step": 175
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7066666666666666,
|
| 573 |
+
"grad_norm": 424.0,
|
| 574 |
+
"learning_rate": 2.3197646927086697e-05,
|
| 575 |
+
"log_odds_chosen": 4.188170433044434,
|
| 576 |
+
"log_odds_ratio": -0.39013969898223877,
|
| 577 |
+
"logps/chosen": -1.56984281539917,
|
| 578 |
+
"logps/rejected": -5.4578375816345215,
|
| 579 |
+
"loss": 14.8288,
|
| 580 |
+
"nll_loss": 1.6269553899765015,
|
| 581 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 582 |
+
"rewards/chosen": -0.784921407699585,
|
| 583 |
+
"rewards/margins": 1.9439977407455444,
|
| 584 |
+
"rewards/rejected": -2.7289187908172607,
|
| 585 |
+
"step": 180
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.7540740740740741,
|
| 589 |
+
"grad_norm": 213.0,
|
| 590 |
+
"learning_rate": 2.1817114703032176e-05,
|
| 591 |
+
"log_odds_chosen": 4.228732585906982,
|
| 592 |
+
"log_odds_ratio": -0.46826067566871643,
|
| 593 |
+
"logps/chosen": -1.6550719738006592,
|
| 594 |
+
"logps/rejected": -5.639611721038818,
|
| 595 |
+
"loss": 15.1508,
|
| 596 |
+
"nll_loss": 1.6106717586517334,
|
| 597 |
+
"rewards/accuracies": 0.809374988079071,
|
| 598 |
+
"rewards/chosen": -0.8275359869003296,
|
| 599 |
+
"rewards/margins": 1.9922701120376587,
|
| 600 |
+
"rewards/rejected": -2.819805860519409,
|
| 601 |
+
"step": 185
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.8014814814814815,
|
| 605 |
+
"grad_norm": 322.0,
|
| 606 |
+
"learning_rate": 2.0446385870993467e-05,
|
| 607 |
+
"log_odds_chosen": 3.5051357746124268,
|
| 608 |
+
"log_odds_ratio": -0.3553612530231476,
|
| 609 |
+
"logps/chosen": -1.289097785949707,
|
| 610 |
+
"logps/rejected": -4.496476650238037,
|
| 611 |
+
"loss": 13.3992,
|
| 612 |
+
"nll_loss": 1.4740468263626099,
|
| 613 |
+
"rewards/accuracies": 0.856249988079071,
|
| 614 |
+
"rewards/chosen": -0.6445488929748535,
|
| 615 |
+
"rewards/margins": 1.603689193725586,
|
| 616 |
+
"rewards/rejected": -2.2482383251190186,
|
| 617 |
+
"step": 190
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.8488888888888888,
|
| 621 |
+
"grad_norm": 440.0,
|
| 622 |
+
"learning_rate": 1.9089682321121834e-05,
|
| 623 |
+
"log_odds_chosen": 4.617656230926514,
|
| 624 |
+
"log_odds_ratio": -0.3456394374370575,
|
| 625 |
+
"logps/chosen": -1.4975018501281738,
|
| 626 |
+
"logps/rejected": -5.828963279724121,
|
| 627 |
+
"loss": 13.1488,
|
| 628 |
+
"nll_loss": 1.6538751125335693,
|
| 629 |
+
"rewards/accuracies": 0.856249988079071,
|
| 630 |
+
"rewards/chosen": -0.7487509250640869,
|
| 631 |
+
"rewards/margins": 2.1657304763793945,
|
| 632 |
+
"rewards/rejected": -2.9144816398620605,
|
| 633 |
+
"step": 195
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.8962962962962964,
|
| 637 |
+
"grad_norm": 516.0,
|
| 638 |
+
"learning_rate": 1.775118274523545e-05,
|
| 639 |
+
"log_odds_chosen": 4.322084426879883,
|
| 640 |
+
"log_odds_ratio": -0.5411938428878784,
|
| 641 |
+
"logps/chosen": -1.8952579498291016,
|
| 642 |
+
"logps/rejected": -5.973208904266357,
|
| 643 |
+
"loss": 21.0282,
|
| 644 |
+
"nll_loss": 1.7580324411392212,
|
| 645 |
+
"rewards/accuracies": 0.809374988079071,
|
| 646 |
+
"rewards/chosen": -0.9476289749145508,
|
| 647 |
+
"rewards/margins": 2.038975477218628,
|
| 648 |
+
"rewards/rejected": -2.9866044521331787,
|
| 649 |
+
"step": 200
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.9437037037037037,
|
| 653 |
+
"grad_norm": 504.0,
|
| 654 |
+
"learning_rate": 1.643500976631037e-05,
|
| 655 |
+
"log_odds_chosen": 3.8178462982177734,
|
| 656 |
+
"log_odds_ratio": -0.41045910120010376,
|
| 657 |
+
"logps/chosen": -1.422102451324463,
|
| 658 |
+
"logps/rejected": -4.953678131103516,
|
| 659 |
+
"loss": 16.8609,
|
| 660 |
+
"nll_loss": 1.5927026271820068,
|
| 661 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 662 |
+
"rewards/chosen": -0.7110512256622314,
|
| 663 |
+
"rewards/margins": 1.7657878398895264,
|
| 664 |
+
"rewards/rejected": -2.476839065551758,
|
| 665 |
+
"step": 205
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.991111111111111,
|
| 669 |
+
"grad_norm": 450.0,
|
| 670 |
+
"learning_rate": 1.514521724066537e-05,
|
| 671 |
+
"log_odds_chosen": 3.2968242168426514,
|
| 672 |
+
"log_odds_ratio": -0.5133435130119324,
|
| 673 |
+
"logps/chosen": -1.5117247104644775,
|
| 674 |
+
"logps/rejected": -4.539895057678223,
|
| 675 |
+
"loss": 19.1746,
|
| 676 |
+
"nll_loss": 1.5904091596603394,
|
| 677 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 678 |
+
"rewards/chosen": -0.7558623552322388,
|
| 679 |
+
"rewards/margins": 1.514085292816162,
|
| 680 |
+
"rewards/rejected": -2.2699475288391113,
|
| 681 |
+
"step": 210
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.0385185185185186,
|
| 685 |
+
"grad_norm": 258.0,
|
| 686 |
+
"learning_rate": 1.3885777771950348e-05,
|
| 687 |
+
"log_odds_chosen": 4.693060398101807,
|
| 688 |
+
"log_odds_ratio": -0.3281119465827942,
|
| 689 |
+
"logps/chosen": -1.2254581451416016,
|
| 690 |
+
"logps/rejected": -5.486399173736572,
|
| 691 |
+
"loss": 2.4338,
|
| 692 |
+
"nll_loss": 1.4177120923995972,
|
| 693 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 694 |
+
"rewards/chosen": -0.6127290725708008,
|
| 695 |
+
"rewards/margins": 2.1304705142974854,
|
| 696 |
+
"rewards/rejected": -2.743199586868286,
|
| 697 |
+
"step": 215
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.0859259259259257,
|
| 701 |
+
"grad_norm": 696.0,
|
| 702 |
+
"learning_rate": 1.2660570475395683e-05,
|
| 703 |
+
"log_odds_chosen": 5.181081295013428,
|
| 704 |
+
"log_odds_ratio": -0.31730157136917114,
|
| 705 |
+
"logps/chosen": -1.451982855796814,
|
| 706 |
+
"logps/rejected": -6.270176887512207,
|
| 707 |
+
"loss": 2.9052,
|
| 708 |
+
"nll_loss": 1.5581401586532593,
|
| 709 |
+
"rewards/accuracies": 0.887499988079071,
|
| 710 |
+
"rewards/chosen": -0.725991427898407,
|
| 711 |
+
"rewards/margins": 2.4090969562530518,
|
| 712 |
+
"rewards/rejected": -3.1350884437561035,
|
| 713 |
+
"step": 220
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.1333333333333333,
|
| 717 |
+
"grad_norm": 294.0,
|
| 718 |
+
"learning_rate": 1.1473369030008974e-05,
|
| 719 |
+
"log_odds_chosen": 6.072526454925537,
|
| 720 |
+
"log_odds_ratio": -0.17935991287231445,
|
| 721 |
+
"logps/chosen": -1.4028112888336182,
|
| 722 |
+
"logps/rejected": -7.085482120513916,
|
| 723 |
+
"loss": -6.7103,
|
| 724 |
+
"nll_loss": 1.5210641622543335,
|
| 725 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 726 |
+
"rewards/chosen": -0.7014056444168091,
|
| 727 |
+
"rewards/margins": 2.8413350582122803,
|
| 728 |
+
"rewards/rejected": -3.542741060256958,
|
| 729 |
+
"step": 225
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.180740740740741,
|
| 733 |
+
"grad_norm": 452.0,
|
| 734 |
+
"learning_rate": 1.0327830055518842e-05,
|
| 735 |
+
"log_odds_chosen": 5.810656547546387,
|
| 736 |
+
"log_odds_ratio": -0.17343996465206146,
|
| 737 |
+
"logps/chosen": -1.257643222808838,
|
| 738 |
+
"logps/rejected": -6.6586809158325195,
|
| 739 |
+
"loss": -6.8616,
|
| 740 |
+
"nll_loss": 1.463210105895996,
|
| 741 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 742 |
+
"rewards/chosen": -0.628821611404419,
|
| 743 |
+
"rewards/margins": 2.70051908493042,
|
| 744 |
+
"rewards/rejected": -3.3293404579162598,
|
| 745 |
+
"step": 230
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.228148148148148,
|
| 749 |
+
"grad_norm": 348.0,
|
| 750 |
+
"learning_rate": 9.227481849865235e-06,
|
| 751 |
+
"log_odds_chosen": 6.33745813369751,
|
| 752 |
+
"log_odds_ratio": -0.2248232066631317,
|
| 753 |
+
"logps/chosen": -1.2892390489578247,
|
| 754 |
+
"logps/rejected": -7.214311122894287,
|
| 755 |
+
"loss": -7.1128,
|
| 756 |
+
"nll_loss": 1.427039384841919,
|
| 757 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 758 |
+
"rewards/chosen": -0.6446195244789124,
|
| 759 |
+
"rewards/margins": 2.962535858154297,
|
| 760 |
+
"rewards/rejected": -3.6071555614471436,
|
| 761 |
+
"step": 235
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 2.2755555555555556,
|
| 765 |
+
"grad_norm": 1248.0,
|
| 766 |
+
"learning_rate": 8.175713521924978e-06,
|
| 767 |
+
"log_odds_chosen": 7.022365570068359,
|
| 768 |
+
"log_odds_ratio": -0.15484580397605896,
|
| 769 |
+
"logps/chosen": -1.3367576599121094,
|
| 770 |
+
"logps/rejected": -7.955646514892578,
|
| 771 |
+
"loss": -10.1801,
|
| 772 |
+
"nll_loss": 1.4348043203353882,
|
| 773 |
+
"rewards/accuracies": 0.934374988079071,
|
| 774 |
+
"rewards/chosen": -0.6683788299560547,
|
| 775 |
+
"rewards/margins": 3.3094444274902344,
|
| 776 |
+
"rewards/rejected": -3.977823257446289,
|
| 777 |
+
"step": 240
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.322962962962963,
|
| 781 |
+
"grad_norm": 640.0,
|
| 782 |
+
"learning_rate": 7.1757645529443665e-06,
|
| 783 |
+
"log_odds_chosen": 7.711653709411621,
|
| 784 |
+
"log_odds_ratio": -0.1372586041688919,
|
| 785 |
+
"logps/chosen": -1.3331178426742554,
|
| 786 |
+
"logps/rejected": -8.624441146850586,
|
| 787 |
+
"loss": -11.5177,
|
| 788 |
+
"nll_loss": 1.4856178760528564,
|
| 789 |
+
"rewards/accuracies": 0.956250011920929,
|
| 790 |
+
"rewards/chosen": -0.6665589213371277,
|
| 791 |
+
"rewards/margins": 3.6456611156463623,
|
| 792 |
+
"rewards/rejected": -4.312220573425293,
|
| 793 |
+
"step": 245
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 2.3703703703703702,
|
| 797 |
+
"grad_norm": 696.0,
|
| 798 |
+
"learning_rate": 6.230714818829733e-06,
|
| 799 |
+
"log_odds_chosen": 7.437263488769531,
|
| 800 |
+
"log_odds_ratio": -0.1762746274471283,
|
| 801 |
+
"logps/chosen": -1.2413113117218018,
|
| 802 |
+
"logps/rejected": -8.23326301574707,
|
| 803 |
+
"loss": -7.4037,
|
| 804 |
+
"nll_loss": 1.4623820781707764,
|
| 805 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 806 |
+
"rewards/chosen": -0.6206556558609009,
|
| 807 |
+
"rewards/margins": 3.495976686477661,
|
| 808 |
+
"rewards/rejected": -4.116631507873535,
|
| 809 |
+
"step": 250
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.417777777777778,
|
| 813 |
+
"grad_norm": 672.0,
|
| 814 |
+
"learning_rate": 5.343475104027743e-06,
|
| 815 |
+
"log_odds_chosen": 7.334293365478516,
|
| 816 |
+
"log_odds_ratio": -0.1758766919374466,
|
| 817 |
+
"logps/chosen": -1.3234690427780151,
|
| 818 |
+
"logps/rejected": -8.235211372375488,
|
| 819 |
+
"loss": -9.7904,
|
| 820 |
+
"nll_loss": 1.4475064277648926,
|
| 821 |
+
"rewards/accuracies": 0.9312499761581421,
|
| 822 |
+
"rewards/chosen": -0.6617345213890076,
|
| 823 |
+
"rewards/margins": 3.4558708667755127,
|
| 824 |
+
"rewards/rejected": -4.117605686187744,
|
| 825 |
+
"step": 255
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.4651851851851854,
|
| 829 |
+
"grad_norm": 624.0,
|
| 830 |
+
"learning_rate": 4.516778136213037e-06,
|
| 831 |
+
"log_odds_chosen": 7.498259544372559,
|
| 832 |
+
"log_odds_ratio": -0.2119707614183426,
|
| 833 |
+
"logps/chosen": -1.427954912185669,
|
| 834 |
+
"logps/rejected": -8.50899600982666,
|
| 835 |
+
"loss": -4.1434,
|
| 836 |
+
"nll_loss": 1.576634407043457,
|
| 837 |
+
"rewards/accuracies": 0.9281250238418579,
|
| 838 |
+
"rewards/chosen": -0.7139774560928345,
|
| 839 |
+
"rewards/margins": 3.540520191192627,
|
| 840 |
+
"rewards/rejected": -4.25449800491333,
|
| 841 |
+
"step": 260
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 2.5125925925925925,
|
| 845 |
+
"grad_norm": 572.0,
|
| 846 |
+
"learning_rate": 3.7531701693965554e-06,
|
| 847 |
+
"log_odds_chosen": 7.056248664855957,
|
| 848 |
+
"log_odds_ratio": -0.2280276119709015,
|
| 849 |
+
"logps/chosen": -1.4004108905792236,
|
| 850 |
+
"logps/rejected": -8.065291404724121,
|
| 851 |
+
"loss": -7.3232,
|
| 852 |
+
"nll_loss": 1.5201685428619385,
|
| 853 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 854 |
+
"rewards/chosen": -0.7002054452896118,
|
| 855 |
+
"rewards/margins": 3.332440137863159,
|
| 856 |
+
"rewards/rejected": -4.0326457023620605,
|
| 857 |
+
"step": 265
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.56,
|
| 861 |
+
"grad_norm": 366.0,
|
| 862 |
+
"learning_rate": 3.055003141378948e-06,
|
| 863 |
+
"log_odds_chosen": 7.216971397399902,
|
| 864 |
+
"log_odds_ratio": -0.21173417568206787,
|
| 865 |
+
"logps/chosen": -1.2760790586471558,
|
| 866 |
+
"logps/rejected": -8.050346374511719,
|
| 867 |
+
"loss": -8.7459,
|
| 868 |
+
"nll_loss": 1.481745958328247,
|
| 869 |
+
"rewards/accuracies": 0.90625,
|
| 870 |
+
"rewards/chosen": -0.6380395293235779,
|
| 871 |
+
"rewards/margins": 3.3871333599090576,
|
| 872 |
+
"rewards/rejected": -4.025173187255859,
|
| 873 |
+
"step": 270
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.6074074074074076,
|
| 877 |
+
"grad_norm": 330.0,
|
| 878 |
+
"learning_rate": 2.424427429704365e-06,
|
| 879 |
+
"log_odds_chosen": 6.806351661682129,
|
| 880 |
+
"log_odds_ratio": -0.17685401439666748,
|
| 881 |
+
"logps/chosen": -1.2406129837036133,
|
| 882 |
+
"logps/rejected": -7.61539363861084,
|
| 883 |
+
"loss": -8.063,
|
| 884 |
+
"nll_loss": 1.4348475933074951,
|
| 885 |
+
"rewards/accuracies": 0.934374988079071,
|
| 886 |
+
"rewards/chosen": -0.6203064918518066,
|
| 887 |
+
"rewards/margins": 3.187389850616455,
|
| 888 |
+
"rewards/rejected": -3.80769681930542,
|
| 889 |
+
"step": 275
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 2.6548148148148147,
|
| 893 |
+
"grad_norm": 800.0,
|
| 894 |
+
"learning_rate": 1.8633852284264508e-06,
|
| 895 |
+
"log_odds_chosen": 6.517434120178223,
|
| 896 |
+
"log_odds_ratio": -0.24552707374095917,
|
| 897 |
+
"logps/chosen": -1.3278234004974365,
|
| 898 |
+
"logps/rejected": -7.444277286529541,
|
| 899 |
+
"loss": -4.5078,
|
| 900 |
+
"nll_loss": 1.491350769996643,
|
| 901 |
+
"rewards/accuracies": 0.887499988079071,
|
| 902 |
+
"rewards/chosen": -0.6639117002487183,
|
| 903 |
+
"rewards/margins": 3.0582268238067627,
|
| 904 |
+
"rewards/rejected": -3.7221386432647705,
|
| 905 |
+
"step": 280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.7022222222222223,
|
| 909 |
+
"grad_norm": 520.0,
|
| 910 |
+
"learning_rate": 1.3736045660864034e-06,
|
| 911 |
+
"log_odds_chosen": 6.652522087097168,
|
| 912 |
+
"log_odds_ratio": -0.22169987857341766,
|
| 913 |
+
"logps/chosen": -1.27732515335083,
|
| 914 |
+
"logps/rejected": -7.5334601402282715,
|
| 915 |
+
"loss": -5.8514,
|
| 916 |
+
"nll_loss": 1.4527100324630737,
|
| 917 |
+
"rewards/accuracies": 0.9125000238418579,
|
| 918 |
+
"rewards/chosen": -0.638662576675415,
|
| 919 |
+
"rewards/margins": 3.1280672550201416,
|
| 920 |
+
"rewards/rejected": -3.7667300701141357,
|
| 921 |
+
"step": 285
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.74962962962963,
|
| 925 |
+
"grad_norm": 470.0,
|
| 926 |
+
"learning_rate": 9.565939833279192e-07,
|
| 927 |
+
"log_odds_chosen": 6.609714508056641,
|
| 928 |
+
"log_odds_ratio": -0.1993839591741562,
|
| 929 |
+
"logps/chosen": -1.4217256307601929,
|
| 930 |
+
"logps/rejected": -7.654360294342041,
|
| 931 |
+
"loss": -3.682,
|
| 932 |
+
"nll_loss": 1.5446010828018188,
|
| 933 |
+
"rewards/accuracies": 0.921875,
|
| 934 |
+
"rewards/chosen": -0.7108628153800964,
|
| 935 |
+
"rewards/margins": 3.1163172721862793,
|
| 936 |
+
"rewards/rejected": -3.8271801471710205,
|
| 937 |
+
"step": 290
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 2.797037037037037,
|
| 941 |
+
"grad_norm": 418.0,
|
| 942 |
+
"learning_rate": 6.136378865420872e-07,
|
| 943 |
+
"log_odds_chosen": 6.339820861816406,
|
| 944 |
+
"log_odds_ratio": -0.2236245572566986,
|
| 945 |
+
"logps/chosen": -1.2482751607894897,
|
| 946 |
+
"logps/rejected": -7.204880714416504,
|
| 947 |
+
"loss": -6.4554,
|
| 948 |
+
"nll_loss": 1.4326238632202148,
|
| 949 |
+
"rewards/accuracies": 0.8843749761581421,
|
| 950 |
+
"rewards/chosen": -0.6241375803947449,
|
| 951 |
+
"rewards/margins": 2.9783027172088623,
|
| 952 |
+
"rewards/rejected": -3.602440357208252,
|
| 953 |
+
"step": 295
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 2.8444444444444446,
|
| 957 |
+
"grad_norm": 478.0,
|
| 958 |
+
"learning_rate": 3.45792591853214e-07,
|
| 959 |
+
"log_odds_chosen": 7.071162223815918,
|
| 960 |
+
"log_odds_ratio": -0.20901203155517578,
|
| 961 |
+
"logps/chosen": -1.270916223526001,
|
| 962 |
+
"logps/rejected": -7.936681270599365,
|
| 963 |
+
"loss": -6.6607,
|
| 964 |
+
"nll_loss": 1.460141897201538,
|
| 965 |
+
"rewards/accuracies": 0.8812500238418579,
|
| 966 |
+
"rewards/chosen": -0.6354581117630005,
|
| 967 |
+
"rewards/margins": 3.3328824043273926,
|
| 968 |
+
"rewards/rejected": -3.9683406352996826,
|
| 969 |
+
"step": 300
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.891851851851852,
|
| 973 |
+
"grad_norm": 410.0,
|
| 974 |
+
"learning_rate": 1.538830716302092e-07,
|
| 975 |
+
"log_odds_chosen": 6.791069030761719,
|
| 976 |
+
"log_odds_ratio": -0.19654326140880585,
|
| 977 |
+
"logps/chosen": -1.2482593059539795,
|
| 978 |
+
"logps/rejected": -7.608553886413574,
|
| 979 |
+
"loss": -7.3947,
|
| 980 |
+
"nll_loss": 1.447270154953003,
|
| 981 |
+
"rewards/accuracies": 0.918749988079071,
|
| 982 |
+
"rewards/chosen": -0.6241296529769897,
|
| 983 |
+
"rewards/margins": 3.180147647857666,
|
| 984 |
+
"rewards/rejected": -3.804276943206787,
|
| 985 |
+
"step": 305
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 2.9392592592592592,
|
| 989 |
+
"grad_norm": 540.0,
|
| 990 |
+
"learning_rate": 3.8500413544415025e-08,
|
| 991 |
+
"log_odds_chosen": 7.128172874450684,
|
| 992 |
+
"log_odds_ratio": -0.1423414945602417,
|
| 993 |
+
"logps/chosen": -1.2163145542144775,
|
| 994 |
+
"logps/rejected": -7.905344486236572,
|
| 995 |
+
"loss": -10.5115,
|
| 996 |
+
"nll_loss": 1.436652421951294,
|
| 997 |
+
"rewards/accuracies": 0.953125,
|
| 998 |
+
"rewards/chosen": -0.6081572771072388,
|
| 999 |
+
"rewards/margins": 3.344515323638916,
|
| 1000 |
+
"rewards/rejected": -3.952672243118286,
|
| 1001 |
+
"step": 310
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 2.986666666666667,
|
| 1005 |
+
"grad_norm": 494.0,
|
| 1006 |
+
"learning_rate": 0.0,
|
| 1007 |
+
"log_odds_chosen": 6.501699924468994,
|
| 1008 |
+
"log_odds_ratio": -0.21002642810344696,
|
| 1009 |
+
"logps/chosen": -1.2352025508880615,
|
| 1010 |
+
"logps/rejected": -7.282286167144775,
|
| 1011 |
+
"loss": -7.439,
|
| 1012 |
+
"nll_loss": 1.396925687789917,
|
| 1013 |
+
"rewards/accuracies": 0.887499988079071,
|
| 1014 |
+
"rewards/chosen": -0.6176012754440308,
|
| 1015 |
+
"rewards/margins": 3.0235419273376465,
|
| 1016 |
+
"rewards/rejected": -3.6411430835723877,
|
| 1017 |
+
"step": 315
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.986666666666667,
|
| 1021 |
+
"step": 315,
|
| 1022 |
+
"total_flos": 0.0,
|
| 1023 |
+
"train_loss": 44.452726506430004,
|
| 1024 |
+
"train_runtime": 7306.7465,
|
| 1025 |
+
"train_samples_per_second": 2.771,
|
| 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 |
+
}
|