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-silvia-basic-5e-5-01-v5
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- orpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for gemma-7b-silvia-basic-5e-5-01-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-silvia-basic-5e-5-01-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/5bupopoy)
|
| 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": 100.64729357522631,
|
| 5 |
+
"train_runtime": 7331.4996,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.762,
|
| 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": 100.64729357522631,
|
| 5 |
+
"train_runtime": 7331.4996,
|
| 6 |
+
"train_samples": 6750,
|
| 7 |
+
"train_samples_per_second": 2.762,
|
| 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": 1472.0,
|
| 14 |
+
"learning_rate": 7.8125e-06,
|
| 15 |
+
"log_odds_chosen": -1.061647653579712,
|
| 16 |
+
"log_odds_ratio": -11.01255989074707,
|
| 17 |
+
"logps/chosen": -22.713951110839844,
|
| 18 |
+
"logps/rejected": -21.652536392211914,
|
| 19 |
+
"loss": 461.0559,
|
| 20 |
+
"nll_loss": 9.837675094604492,
|
| 21 |
+
"rewards/accuracies": 0.484375,
|
| 22 |
+
"rewards/chosen": -2.271395206451416,
|
| 23 |
+
"rewards/margins": -0.10614162683486938,
|
| 24 |
+
"rewards/rejected": -2.1652536392211914,
|
| 25 |
+
"step": 5
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.09481481481481481,
|
| 29 |
+
"grad_norm": 1000.0,
|
| 30 |
+
"learning_rate": 1.5625e-05,
|
| 31 |
+
"log_odds_chosen": -3.1320607662200928,
|
| 32 |
+
"log_odds_ratio": -11.077211380004883,
|
| 33 |
+
"logps/chosen": -22.03810691833496,
|
| 34 |
+
"logps/rejected": -18.906375885009766,
|
| 35 |
+
"loss": 576.5301,
|
| 36 |
+
"nll_loss": 8.74990177154541,
|
| 37 |
+
"rewards/accuracies": 0.4468750059604645,
|
| 38 |
+
"rewards/chosen": -2.203810453414917,
|
| 39 |
+
"rewards/margins": -0.3131731152534485,
|
| 40 |
+
"rewards/rejected": -1.8906371593475342,
|
| 41 |
+
"step": 10
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.14222222222222222,
|
| 45 |
+
"grad_norm": 876.0,
|
| 46 |
+
"learning_rate": 2.34375e-05,
|
| 47 |
+
"log_odds_chosen": -1.2847745418548584,
|
| 48 |
+
"log_odds_ratio": -11.418578147888184,
|
| 49 |
+
"logps/chosen": -22.122282028198242,
|
| 50 |
+
"logps/rejected": -20.836055755615234,
|
| 51 |
+
"loss": 658.1847,
|
| 52 |
+
"nll_loss": 8.011680603027344,
|
| 53 |
+
"rewards/accuracies": 0.515625,
|
| 54 |
+
"rewards/chosen": -2.212228298187256,
|
| 55 |
+
"rewards/margins": -0.12862251698970795,
|
| 56 |
+
"rewards/rejected": -2.0836057662963867,
|
| 57 |
+
"step": 15
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.18962962962962962,
|
| 61 |
+
"grad_norm": 5696.0,
|
| 62 |
+
"learning_rate": 3.125e-05,
|
| 63 |
+
"log_odds_chosen": -4.7826385498046875,
|
| 64 |
+
"log_odds_ratio": -10.32951831817627,
|
| 65 |
+
"logps/chosen": -18.428646087646484,
|
| 66 |
+
"logps/rejected": -13.647994995117188,
|
| 67 |
+
"loss": 619.2777,
|
| 68 |
+
"nll_loss": 6.516907691955566,
|
| 69 |
+
"rewards/accuracies": 0.4312500059604645,
|
| 70 |
+
"rewards/chosen": -1.842864751815796,
|
| 71 |
+
"rewards/margins": -0.47806549072265625,
|
| 72 |
+
"rewards/rejected": -1.3647992610931396,
|
| 73 |
+
"step": 20
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.23703703703703705,
|
| 77 |
+
"grad_norm": 1440.0,
|
| 78 |
+
"learning_rate": 3.90625e-05,
|
| 79 |
+
"log_odds_chosen": -0.21342229843139648,
|
| 80 |
+
"log_odds_ratio": -2.4648327827453613,
|
| 81 |
+
"logps/chosen": -5.189468860626221,
|
| 82 |
+
"logps/rejected": -4.9640350341796875,
|
| 83 |
+
"loss": 238.5154,
|
| 84 |
+
"nll_loss": 2.8294453620910645,
|
| 85 |
+
"rewards/accuracies": 0.534375011920929,
|
| 86 |
+
"rewards/chosen": -0.5189469456672668,
|
| 87 |
+
"rewards/margins": -0.022543441504240036,
|
| 88 |
+
"rewards/rejected": -0.49640339612960815,
|
| 89 |
+
"step": 25
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.28444444444444444,
|
| 93 |
+
"grad_norm": 600.0,
|
| 94 |
+
"learning_rate": 4.6875e-05,
|
| 95 |
+
"log_odds_chosen": 0.027130698785185814,
|
| 96 |
+
"log_odds_ratio": -0.9074820280075073,
|
| 97 |
+
"logps/chosen": -1.7969779968261719,
|
| 98 |
+
"logps/rejected": -1.825482726097107,
|
| 99 |
+
"loss": 143.8757,
|
| 100 |
+
"nll_loss": 1.9630838632583618,
|
| 101 |
+
"rewards/accuracies": 0.5218750238418579,
|
| 102 |
+
"rewards/chosen": -0.17969779670238495,
|
| 103 |
+
"rewards/margins": 0.002850478980690241,
|
| 104 |
+
"rewards/rejected": -0.18254828453063965,
|
| 105 |
+
"step": 30
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.33185185185185184,
|
| 109 |
+
"grad_norm": 1304.0,
|
| 110 |
+
"learning_rate": 4.998613757348784e-05,
|
| 111 |
+
"log_odds_chosen": 0.19788822531700134,
|
| 112 |
+
"log_odds_ratio": -0.8488085865974426,
|
| 113 |
+
"logps/chosen": -1.7060260772705078,
|
| 114 |
+
"logps/rejected": -1.8849117755889893,
|
| 115 |
+
"loss": 132.2157,
|
| 116 |
+
"nll_loss": 1.8448913097381592,
|
| 117 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 118 |
+
"rewards/chosen": -0.17060261964797974,
|
| 119 |
+
"rewards/margins": 0.017888568341732025,
|
| 120 |
+
"rewards/rejected": -0.18849118053913116,
|
| 121 |
+
"step": 35
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.37925925925925924,
|
| 125 |
+
"grad_norm": 446.0,
|
| 126 |
+
"learning_rate": 4.990147841143462e-05,
|
| 127 |
+
"log_odds_chosen": 0.23903241753578186,
|
| 128 |
+
"log_odds_ratio": -0.7467631697654724,
|
| 129 |
+
"logps/chosen": -1.4654309749603271,
|
| 130 |
+
"logps/rejected": -1.6631317138671875,
|
| 131 |
+
"loss": 123.9697,
|
| 132 |
+
"nll_loss": 1.7629951238632202,
|
| 133 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 134 |
+
"rewards/chosen": -0.14654308557510376,
|
| 135 |
+
"rewards/margins": 0.019770082086324692,
|
| 136 |
+
"rewards/rejected": -0.16631318628787994,
|
| 137 |
+
"step": 40
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.4266666666666667,
|
| 141 |
+
"grad_norm": 292.0,
|
| 142 |
+
"learning_rate": 4.97401218720448e-05,
|
| 143 |
+
"log_odds_chosen": 0.18752680718898773,
|
| 144 |
+
"log_odds_ratio": -0.7384530901908875,
|
| 145 |
+
"logps/chosen": -1.374541997909546,
|
| 146 |
+
"logps/rejected": -1.5379812717437744,
|
| 147 |
+
"loss": 114.1024,
|
| 148 |
+
"nll_loss": 1.662832498550415,
|
| 149 |
+
"rewards/accuracies": 0.565625011920929,
|
| 150 |
+
"rewards/chosen": -0.13745419681072235,
|
| 151 |
+
"rewards/margins": 0.016343925148248672,
|
| 152 |
+
"rewards/rejected": -0.15379810333251953,
|
| 153 |
+
"step": 45
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.4740740740740741,
|
| 157 |
+
"grad_norm": 434.0,
|
| 158 |
+
"learning_rate": 4.9502564938797946e-05,
|
| 159 |
+
"log_odds_chosen": 0.24252310395240784,
|
| 160 |
+
"log_odds_ratio": -0.7051388025283813,
|
| 161 |
+
"logps/chosen": -1.2654331922531128,
|
| 162 |
+
"logps/rejected": -1.4687267541885376,
|
| 163 |
+
"loss": 102.5071,
|
| 164 |
+
"nll_loss": 1.5553805828094482,
|
| 165 |
+
"rewards/accuracies": 0.590624988079071,
|
| 166 |
+
"rewards/chosen": -0.126543328166008,
|
| 167 |
+
"rewards/margins": 0.020329322665929794,
|
| 168 |
+
"rewards/rejected": -0.1468726545572281,
|
| 169 |
+
"step": 50
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.5214814814814814,
|
| 173 |
+
"grad_norm": 524.0,
|
| 174 |
+
"learning_rate": 4.918953929490768e-05,
|
| 175 |
+
"log_odds_chosen": 0.20942315459251404,
|
| 176 |
+
"log_odds_ratio": -0.7282800078392029,
|
| 177 |
+
"logps/chosen": -1.2852176427841187,
|
| 178 |
+
"logps/rejected": -1.4517309665679932,
|
| 179 |
+
"loss": 106.6771,
|
| 180 |
+
"nll_loss": 1.5853464603424072,
|
| 181 |
+
"rewards/accuracies": 0.581250011920929,
|
| 182 |
+
"rewards/chosen": -0.12852177023887634,
|
| 183 |
+
"rewards/margins": 0.016651321202516556,
|
| 184 |
+
"rewards/rejected": -0.1451731026172638,
|
| 185 |
+
"step": 55
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.5688888888888889,
|
| 189 |
+
"grad_norm": 318.0,
|
| 190 |
+
"learning_rate": 4.88020090697132e-05,
|
| 191 |
+
"log_odds_chosen": 0.17723873257637024,
|
| 192 |
+
"log_odds_ratio": -0.7213740348815918,
|
| 193 |
+
"logps/chosen": -1.2767646312713623,
|
| 194 |
+
"logps/rejected": -1.4264848232269287,
|
| 195 |
+
"loss": 107.5658,
|
| 196 |
+
"nll_loss": 1.5910738706588745,
|
| 197 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 198 |
+
"rewards/chosen": -0.12767647206783295,
|
| 199 |
+
"rewards/margins": 0.014972020871937275,
|
| 200 |
+
"rewards/rejected": -0.14264848828315735,
|
| 201 |
+
"step": 60
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.6162962962962963,
|
| 205 |
+
"grad_norm": 656.0,
|
| 206 |
+
"learning_rate": 4.834116786912897e-05,
|
| 207 |
+
"log_odds_chosen": 0.22847473621368408,
|
| 208 |
+
"log_odds_ratio": -0.6963309645652771,
|
| 209 |
+
"logps/chosen": -1.2379462718963623,
|
| 210 |
+
"logps/rejected": -1.4141137599945068,
|
| 211 |
+
"loss": 100.0757,
|
| 212 |
+
"nll_loss": 1.5201084613800049,
|
| 213 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 214 |
+
"rewards/chosen": -0.12379463016986847,
|
| 215 |
+
"rewards/margins": 0.01761675253510475,
|
| 216 |
+
"rewards/rejected": -0.14141139388084412,
|
| 217 |
+
"step": 65
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.6637037037037037,
|
| 221 |
+
"grad_norm": 624.0,
|
| 222 |
+
"learning_rate": 4.7808435099299045e-05,
|
| 223 |
+
"log_odds_chosen": 0.3744918704032898,
|
| 224 |
+
"log_odds_ratio": -0.6574251651763916,
|
| 225 |
+
"logps/chosen": -1.2112077474594116,
|
| 226 |
+
"logps/rejected": -1.520322561264038,
|
| 227 |
+
"loss": 99.2833,
|
| 228 |
+
"nll_loss": 1.510595679283142,
|
| 229 |
+
"rewards/accuracies": 0.628125011920929,
|
| 230 |
+
"rewards/chosen": -0.12112077325582504,
|
| 231 |
+
"rewards/margins": 0.030911466106772423,
|
| 232 |
+
"rewards/rejected": -0.15203224122524261,
|
| 233 |
+
"step": 70
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.7111111111111111,
|
| 237 |
+
"grad_norm": 612.0,
|
| 238 |
+
"learning_rate": 4.720545159477922e-05,
|
| 239 |
+
"log_odds_chosen": 0.29794445633888245,
|
| 240 |
+
"log_odds_ratio": -0.7010519504547119,
|
| 241 |
+
"logps/chosen": -1.2030102014541626,
|
| 242 |
+
"logps/rejected": -1.437780499458313,
|
| 243 |
+
"loss": 98.7744,
|
| 244 |
+
"nll_loss": 1.4964594841003418,
|
| 245 |
+
"rewards/accuracies": 0.6156250238418579,
|
| 246 |
+
"rewards/chosen": -0.1203010231256485,
|
| 247 |
+
"rewards/margins": 0.023477038368582726,
|
| 248 |
+
"rewards/rejected": -0.14377805590629578,
|
| 249 |
+
"step": 75
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.7585185185185185,
|
| 253 |
+
"grad_norm": 510.0,
|
| 254 |
+
"learning_rate": 4.653407456471222e-05,
|
| 255 |
+
"log_odds_chosen": 0.19871729612350464,
|
| 256 |
+
"log_odds_ratio": -0.6991140842437744,
|
| 257 |
+
"logps/chosen": -1.1551485061645508,
|
| 258 |
+
"logps/rejected": -1.3054516315460205,
|
| 259 |
+
"loss": 92.3435,
|
| 260 |
+
"nll_loss": 1.4383039474487305,
|
| 261 |
+
"rewards/accuracies": 0.550000011920929,
|
| 262 |
+
"rewards/chosen": -0.1155148595571518,
|
| 263 |
+
"rewards/margins": 0.015030312351882458,
|
| 264 |
+
"rewards/rejected": -0.13054516911506653,
|
| 265 |
+
"step": 80
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 0.8059259259259259,
|
| 269 |
+
"grad_norm": 344.0,
|
| 270 |
+
"learning_rate": 4.579637187256222e-05,
|
| 271 |
+
"log_odds_chosen": 0.21253113448619843,
|
| 272 |
+
"log_odds_ratio": -0.6881605386734009,
|
| 273 |
+
"logps/chosen": -1.144791841506958,
|
| 274 |
+
"logps/rejected": -1.3137983083724976,
|
| 275 |
+
"loss": 97.0547,
|
| 276 |
+
"nll_loss": 1.4676554203033447,
|
| 277 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 278 |
+
"rewards/chosen": -0.1144791841506958,
|
| 279 |
+
"rewards/margins": 0.016900673508644104,
|
| 280 |
+
"rewards/rejected": -0.1313798427581787,
|
| 281 |
+
"step": 85
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.8533333333333334,
|
| 285 |
+
"grad_norm": 180.0,
|
| 286 |
+
"learning_rate": 4.499461566702685e-05,
|
| 287 |
+
"log_odds_chosen": 0.2872636616230011,
|
| 288 |
+
"log_odds_ratio": -0.694359302520752,
|
| 289 |
+
"logps/chosen": -1.0993402004241943,
|
| 290 |
+
"logps/rejected": -1.3027684688568115,
|
| 291 |
+
"loss": 89.232,
|
| 292 |
+
"nll_loss": 1.395646095275879,
|
| 293 |
+
"rewards/accuracies": 0.637499988079071,
|
| 294 |
+
"rewards/chosen": -0.1099340170621872,
|
| 295 |
+
"rewards/margins": 0.020342819392681122,
|
| 296 |
+
"rewards/rejected": -0.13027682900428772,
|
| 297 |
+
"step": 90
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.9007407407407407,
|
| 301 |
+
"grad_norm": 444.0,
|
| 302 |
+
"learning_rate": 4.413127538374411e-05,
|
| 303 |
+
"log_odds_chosen": 0.15917561948299408,
|
| 304 |
+
"log_odds_ratio": -0.7126652598381042,
|
| 305 |
+
"logps/chosen": -1.129868745803833,
|
| 306 |
+
"logps/rejected": -1.2550439834594727,
|
| 307 |
+
"loss": 96.0778,
|
| 308 |
+
"nll_loss": 1.448910117149353,
|
| 309 |
+
"rewards/accuracies": 0.5406249761581421,
|
| 310 |
+
"rewards/chosen": -0.11298687756061554,
|
| 311 |
+
"rewards/margins": 0.012517519295215607,
|
| 312 |
+
"rewards/rejected": -0.12550440430641174,
|
| 313 |
+
"step": 95
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.9481481481481482,
|
| 317 |
+
"grad_norm": 412.0,
|
| 318 |
+
"learning_rate": 4.320901013934887e-05,
|
| 319 |
+
"log_odds_chosen": 0.22853884100914001,
|
| 320 |
+
"log_odds_ratio": -0.6652454137802124,
|
| 321 |
+
"logps/chosen": -1.093160629272461,
|
| 322 |
+
"logps/rejected": -1.260899305343628,
|
| 323 |
+
"loss": 90.0544,
|
| 324 |
+
"nll_loss": 1.4013230800628662,
|
| 325 |
+
"rewards/accuracies": 0.596875011920929,
|
| 326 |
+
"rewards/chosen": -0.10931606590747833,
|
| 327 |
+
"rewards/margins": 0.01677386835217476,
|
| 328 |
+
"rewards/rejected": -0.1260899305343628,
|
| 329 |
+
"step": 100
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.9955555555555555,
|
| 333 |
+
"grad_norm": 330.0,
|
| 334 |
+
"learning_rate": 4.223066054130568e-05,
|
| 335 |
+
"log_odds_chosen": 0.28896692395210266,
|
| 336 |
+
"log_odds_ratio": -0.6708310842514038,
|
| 337 |
+
"logps/chosen": -1.0797439813613892,
|
| 338 |
+
"logps/rejected": -1.315000057220459,
|
| 339 |
+
"loss": 93.9254,
|
| 340 |
+
"nll_loss": 1.4331862926483154,
|
| 341 |
+
"rewards/accuracies": 0.59375,
|
| 342 |
+
"rewards/chosen": -0.10797439515590668,
|
| 343 |
+
"rewards/margins": 0.02352561056613922,
|
| 344 |
+
"rewards/rejected": -0.1315000057220459,
|
| 345 |
+
"step": 105
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.0429629629629629,
|
| 349 |
+
"grad_norm": 204.0,
|
| 350 |
+
"learning_rate": 4.1199239938743797e-05,
|
| 351 |
+
"log_odds_chosen": 0.569130003452301,
|
| 352 |
+
"log_odds_ratio": -0.5484132170677185,
|
| 353 |
+
"logps/chosen": -0.9365727305412292,
|
| 354 |
+
"logps/rejected": -1.3183672428131104,
|
| 355 |
+
"loss": 66.1124,
|
| 356 |
+
"nll_loss": 1.20535147190094,
|
| 357 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 358 |
+
"rewards/chosen": -0.09365727007389069,
|
| 359 |
+
"rewards/margins": 0.03817944601178169,
|
| 360 |
+
"rewards/rejected": -0.13183671236038208,
|
| 361 |
+
"step": 110
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.0903703703703704,
|
| 365 |
+
"grad_norm": 227.0,
|
| 366 |
+
"learning_rate": 4.0117925141242174e-05,
|
| 367 |
+
"log_odds_chosen": 0.7776085138320923,
|
| 368 |
+
"log_odds_ratio": -0.5151402950286865,
|
| 369 |
+
"logps/chosen": -0.8830581903457642,
|
| 370 |
+
"logps/rejected": -1.399554967880249,
|
| 371 |
+
"loss": 63.4462,
|
| 372 |
+
"nll_loss": 1.1780006885528564,
|
| 373 |
+
"rewards/accuracies": 0.737500011920929,
|
| 374 |
+
"rewards/chosen": -0.08830581605434418,
|
| 375 |
+
"rewards/margins": 0.051649678498506546,
|
| 376 |
+
"rewards/rejected": -0.13995549082756042,
|
| 377 |
+
"step": 115
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.1377777777777778,
|
| 381 |
+
"grad_norm": 171.0,
|
| 382 |
+
"learning_rate": 3.899004663415084e-05,
|
| 383 |
+
"log_odds_chosen": 0.6698681712150574,
|
| 384 |
+
"log_odds_ratio": -0.5222538113594055,
|
| 385 |
+
"logps/chosen": -0.8659135103225708,
|
| 386 |
+
"logps/rejected": -1.3088980913162231,
|
| 387 |
+
"loss": 64.559,
|
| 388 |
+
"nll_loss": 1.1794170141220093,
|
| 389 |
+
"rewards/accuracies": 0.731249988079071,
|
| 390 |
+
"rewards/chosen": -0.08659134805202484,
|
| 391 |
+
"rewards/margins": 0.0442984513938427,
|
| 392 |
+
"rewards/rejected": -0.13088980317115784,
|
| 393 |
+
"step": 120
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.1851851851851851,
|
| 397 |
+
"grad_norm": 262.0,
|
| 398 |
+
"learning_rate": 3.781907832058587e-05,
|
| 399 |
+
"log_odds_chosen": 0.6859734654426575,
|
| 400 |
+
"log_odds_ratio": -0.5173711776733398,
|
| 401 |
+
"logps/chosen": -0.9302700161933899,
|
| 402 |
+
"logps/rejected": -1.3908751010894775,
|
| 403 |
+
"loss": 62.7453,
|
| 404 |
+
"nll_loss": 1.1762224435806274,
|
| 405 |
+
"rewards/accuracies": 0.7406250238418579,
|
| 406 |
+
"rewards/chosen": -0.0930270105600357,
|
| 407 |
+
"rewards/margins": 0.046060509979724884,
|
| 408 |
+
"rewards/rejected": -0.1390875279903412,
|
| 409 |
+
"step": 125
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.2325925925925927,
|
| 413 |
+
"grad_norm": 170.0,
|
| 414 |
+
"learning_rate": 3.660862682169282e-05,
|
| 415 |
+
"log_odds_chosen": 0.795188307762146,
|
| 416 |
+
"log_odds_ratio": -0.5165600776672363,
|
| 417 |
+
"logps/chosen": -0.8818154335021973,
|
| 418 |
+
"logps/rejected": -1.420803189277649,
|
| 419 |
+
"loss": 61.3166,
|
| 420 |
+
"nll_loss": 1.1659226417541504,
|
| 421 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 422 |
+
"rewards/chosen": -0.08818155527114868,
|
| 423 |
+
"rewards/margins": 0.05389877036213875,
|
| 424 |
+
"rewards/rejected": -0.14208030700683594,
|
| 425 |
+
"step": 130
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.28,
|
| 429 |
+
"grad_norm": 216.0,
|
| 430 |
+
"learning_rate": 3.5362420368134356e-05,
|
| 431 |
+
"log_odds_chosen": 0.7150818109512329,
|
| 432 |
+
"log_odds_ratio": -0.5056425333023071,
|
| 433 |
+
"logps/chosen": -0.8704392313957214,
|
| 434 |
+
"logps/rejected": -1.3260860443115234,
|
| 435 |
+
"loss": 62.9974,
|
| 436 |
+
"nll_loss": 1.165181279182434,
|
| 437 |
+
"rewards/accuracies": 0.75,
|
| 438 |
+
"rewards/chosen": -0.08704392611980438,
|
| 439 |
+
"rewards/margins": 0.0455646738409996,
|
| 440 |
+
"rewards/rejected": -0.13260860741138458,
|
| 441 |
+
"step": 135
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.3274074074074074,
|
| 445 |
+
"grad_norm": 216.0,
|
| 446 |
+
"learning_rate": 3.408429731701635e-05,
|
| 447 |
+
"log_odds_chosen": 0.8086262941360474,
|
| 448 |
+
"log_odds_ratio": -0.5228279829025269,
|
| 449 |
+
"logps/chosen": -0.8993717432022095,
|
| 450 |
+
"logps/rejected": -1.4720028638839722,
|
| 451 |
+
"loss": 66.0169,
|
| 452 |
+
"nll_loss": 1.1864889860153198,
|
| 453 |
+
"rewards/accuracies": 0.746874988079071,
|
| 454 |
+
"rewards/chosen": -0.08993717283010483,
|
| 455 |
+
"rewards/margins": 0.057263124734163284,
|
| 456 |
+
"rewards/rejected": -0.1472003012895584,
|
| 457 |
+
"step": 140
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.374814814814815,
|
| 461 |
+
"grad_norm": 209.0,
|
| 462 |
+
"learning_rate": 3.2778194329621104e-05,
|
| 463 |
+
"log_odds_chosen": 0.6819108724594116,
|
| 464 |
+
"log_odds_ratio": -0.5213868021965027,
|
| 465 |
+
"logps/chosen": -0.9046201705932617,
|
| 466 |
+
"logps/rejected": -1.3656418323516846,
|
| 467 |
+
"loss": 66.4385,
|
| 468 |
+
"nll_loss": 1.199620008468628,
|
| 469 |
+
"rewards/accuracies": 0.7406250238418579,
|
| 470 |
+
"rewards/chosen": -0.09046201407909393,
|
| 471 |
+
"rewards/margins": 0.04610215872526169,
|
| 472 |
+
"rewards/rejected": -0.13656418025493622,
|
| 473 |
+
"step": 145
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.4222222222222223,
|
| 477 |
+
"grad_norm": 268.0,
|
| 478 |
+
"learning_rate": 3.144813424636031e-05,
|
| 479 |
+
"log_odds_chosen": 0.7690508961677551,
|
| 480 |
+
"log_odds_ratio": -0.5032103657722473,
|
| 481 |
+
"logps/chosen": -0.8478023409843445,
|
| 482 |
+
"logps/rejected": -1.3725353479385376,
|
| 483 |
+
"loss": 62.1945,
|
| 484 |
+
"nll_loss": 1.1453306674957275,
|
| 485 |
+
"rewards/accuracies": 0.731249988079071,
|
| 486 |
+
"rewards/chosen": -0.08478023111820221,
|
| 487 |
+
"rewards/margins": 0.052473314106464386,
|
| 488 |
+
"rewards/rejected": -0.137253537774086,
|
| 489 |
+
"step": 150
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.4696296296296296,
|
| 493 |
+
"grad_norm": 176.0,
|
| 494 |
+
"learning_rate": 3.0098213696293542e-05,
|
| 495 |
+
"log_odds_chosen": 0.8440017700195312,
|
| 496 |
+
"log_odds_ratio": -0.5293614268302917,
|
| 497 |
+
"logps/chosen": -0.8639848828315735,
|
| 498 |
+
"logps/rejected": -1.490368127822876,
|
| 499 |
+
"loss": 63.6013,
|
| 500 |
+
"nll_loss": 1.1569997072219849,
|
| 501 |
+
"rewards/accuracies": 0.737500011920929,
|
| 502 |
+
"rewards/chosen": -0.08639849722385406,
|
| 503 |
+
"rewards/margins": 0.06263832747936249,
|
| 504 |
+
"rewards/rejected": -0.14903683960437775,
|
| 505 |
+
"step": 155
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 1.5170370370370372,
|
| 509 |
+
"grad_norm": 170.0,
|
| 510 |
+
"learning_rate": 2.8732590479375165e-05,
|
| 511 |
+
"log_odds_chosen": 0.750348687171936,
|
| 512 |
+
"log_odds_ratio": -0.544133186340332,
|
| 513 |
+
"logps/chosen": -0.8639837503433228,
|
| 514 |
+
"logps/rejected": -1.382258415222168,
|
| 515 |
+
"loss": 60.6858,
|
| 516 |
+
"nll_loss": 1.1391737461090088,
|
| 517 |
+
"rewards/accuracies": 0.699999988079071,
|
| 518 |
+
"rewards/chosen": -0.08639837801456451,
|
| 519 |
+
"rewards/margins": 0.051827479153871536,
|
| 520 |
+
"rewards/rejected": -0.13822586834430695,
|
| 521 |
+
"step": 160
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.5644444444444443,
|
| 525 |
+
"grad_norm": 282.0,
|
| 526 |
+
"learning_rate": 2.7355470760292956e-05,
|
| 527 |
+
"log_odds_chosen": 0.981040358543396,
|
| 528 |
+
"log_odds_ratio": -0.4906982481479645,
|
| 529 |
+
"logps/chosen": -0.8270002603530884,
|
| 530 |
+
"logps/rejected": -1.5326000452041626,
|
| 531 |
+
"loss": 63.172,
|
| 532 |
+
"nll_loss": 1.1677793264389038,
|
| 533 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 534 |
+
"rewards/chosen": -0.08270002901554108,
|
| 535 |
+
"rewards/margins": 0.07055998593568802,
|
| 536 |
+
"rewards/rejected": -0.1532600075006485,
|
| 537 |
+
"step": 165
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.6118518518518519,
|
| 541 |
+
"grad_norm": 364.0,
|
| 542 |
+
"learning_rate": 2.597109611334169e-05,
|
| 543 |
+
"log_odds_chosen": 0.8211015462875366,
|
| 544 |
+
"log_odds_ratio": -0.516723096370697,
|
| 545 |
+
"logps/chosen": -0.8916131258010864,
|
| 546 |
+
"logps/rejected": -1.4886616468429565,
|
| 547 |
+
"loss": 61.9588,
|
| 548 |
+
"nll_loss": 1.1598554849624634,
|
| 549 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 550 |
+
"rewards/chosen": -0.08916130661964417,
|
| 551 |
+
"rewards/margins": 0.05970486253499985,
|
| 552 |
+
"rewards/rejected": -0.1488661766052246,
|
| 553 |
+
"step": 170
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.6592592592592592,
|
| 557 |
+
"grad_norm": 177.0,
|
| 558 |
+
"learning_rate": 2.458373045823404e-05,
|
| 559 |
+
"log_odds_chosen": 0.8482476472854614,
|
| 560 |
+
"log_odds_ratio": -0.49223431944847107,
|
| 561 |
+
"logps/chosen": -0.8441880345344543,
|
| 562 |
+
"logps/rejected": -1.4245407581329346,
|
| 563 |
+
"loss": 58.6637,
|
| 564 |
+
"nll_loss": 1.1308585405349731,
|
| 565 |
+
"rewards/accuracies": 0.753125011920929,
|
| 566 |
+
"rewards/chosen": -0.08441879600286484,
|
| 567 |
+
"rewards/margins": 0.05803527683019638,
|
| 568 |
+
"rewards/rejected": -0.1424541026353836,
|
| 569 |
+
"step": 175
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7066666666666666,
|
| 573 |
+
"grad_norm": 294.0,
|
| 574 |
+
"learning_rate": 2.3197646927086697e-05,
|
| 575 |
+
"log_odds_chosen": 0.8345856666564941,
|
| 576 |
+
"log_odds_ratio": -0.5028634667396545,
|
| 577 |
+
"logps/chosen": -0.8489413261413574,
|
| 578 |
+
"logps/rejected": -1.4371174573898315,
|
| 579 |
+
"loss": 58.5978,
|
| 580 |
+
"nll_loss": 1.133784294128418,
|
| 581 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 582 |
+
"rewards/chosen": -0.08489413559436798,
|
| 583 |
+
"rewards/margins": 0.05881761386990547,
|
| 584 |
+
"rewards/rejected": -0.14371174573898315,
|
| 585 |
+
"step": 180
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 1.7540740740740741,
|
| 589 |
+
"grad_norm": 234.0,
|
| 590 |
+
"learning_rate": 2.1817114703032176e-05,
|
| 591 |
+
"log_odds_chosen": 0.8239284753799438,
|
| 592 |
+
"log_odds_ratio": -0.5574834942817688,
|
| 593 |
+
"logps/chosen": -0.9100497961044312,
|
| 594 |
+
"logps/rejected": -1.5430285930633545,
|
| 595 |
+
"loss": 63.5627,
|
| 596 |
+
"nll_loss": 1.1603102684020996,
|
| 597 |
+
"rewards/accuracies": 0.703125,
|
| 598 |
+
"rewards/chosen": -0.09100499004125595,
|
| 599 |
+
"rewards/margins": 0.06329786032438278,
|
| 600 |
+
"rewards/rejected": -0.15430285036563873,
|
| 601 |
+
"step": 185
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.8014814814814815,
|
| 605 |
+
"grad_norm": 157.0,
|
| 606 |
+
"learning_rate": 2.0446385870993467e-05,
|
| 607 |
+
"log_odds_chosen": 0.9324125051498413,
|
| 608 |
+
"log_odds_ratio": -0.5415812730789185,
|
| 609 |
+
"logps/chosen": -0.894865870475769,
|
| 610 |
+
"logps/rejected": -1.5854549407958984,
|
| 611 |
+
"loss": 63.374,
|
| 612 |
+
"nll_loss": 1.1591455936431885,
|
| 613 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 614 |
+
"rewards/chosen": -0.08948659896850586,
|
| 615 |
+
"rewards/margins": 0.06905890256166458,
|
| 616 |
+
"rewards/rejected": -0.15854549407958984,
|
| 617 |
+
"step": 190
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.8488888888888888,
|
| 621 |
+
"grad_norm": 197.0,
|
| 622 |
+
"learning_rate": 1.9089682321121834e-05,
|
| 623 |
+
"log_odds_chosen": 1.1004116535186768,
|
| 624 |
+
"log_odds_ratio": -0.4837014079093933,
|
| 625 |
+
"logps/chosen": -0.8859478235244751,
|
| 626 |
+
"logps/rejected": -1.7115284204483032,
|
| 627 |
+
"loss": 61.1523,
|
| 628 |
+
"nll_loss": 1.149214506149292,
|
| 629 |
+
"rewards/accuracies": 0.765625,
|
| 630 |
+
"rewards/chosen": -0.08859477937221527,
|
| 631 |
+
"rewards/margins": 0.0825580507516861,
|
| 632 |
+
"rewards/rejected": -0.17115284502506256,
|
| 633 |
+
"step": 195
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.8962962962962964,
|
| 637 |
+
"grad_norm": 282.0,
|
| 638 |
+
"learning_rate": 1.775118274523545e-05,
|
| 639 |
+
"log_odds_chosen": 0.9893839955329895,
|
| 640 |
+
"log_odds_ratio": -0.5392941832542419,
|
| 641 |
+
"logps/chosen": -0.9068052172660828,
|
| 642 |
+
"logps/rejected": -1.6685311794281006,
|
| 643 |
+
"loss": 63.8627,
|
| 644 |
+
"nll_loss": 1.1626369953155518,
|
| 645 |
+
"rewards/accuracies": 0.690625011920929,
|
| 646 |
+
"rewards/chosen": -0.09068052470684052,
|
| 647 |
+
"rewards/margins": 0.0761725977063179,
|
| 648 |
+
"rewards/rejected": -0.1668531447649002,
|
| 649 |
+
"step": 200
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.9437037037037037,
|
| 653 |
+
"grad_norm": 256.0,
|
| 654 |
+
"learning_rate": 1.643500976631037e-05,
|
| 655 |
+
"log_odds_chosen": 0.9225565195083618,
|
| 656 |
+
"log_odds_ratio": -0.5391720533370972,
|
| 657 |
+
"logps/chosen": -0.8900644183158875,
|
| 658 |
+
"logps/rejected": -1.5420877933502197,
|
| 659 |
+
"loss": 63.8396,
|
| 660 |
+
"nll_loss": 1.1569474935531616,
|
| 661 |
+
"rewards/accuracies": 0.7406250238418579,
|
| 662 |
+
"rewards/chosen": -0.0890064537525177,
|
| 663 |
+
"rewards/margins": 0.06520231813192368,
|
| 664 |
+
"rewards/rejected": -0.15420876443386078,
|
| 665 |
+
"step": 205
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.991111111111111,
|
| 669 |
+
"grad_norm": 209.0,
|
| 670 |
+
"learning_rate": 1.514521724066537e-05,
|
| 671 |
+
"log_odds_chosen": 0.7736507654190063,
|
| 672 |
+
"log_odds_ratio": -0.5372087955474854,
|
| 673 |
+
"logps/chosen": -0.8860548734664917,
|
| 674 |
+
"logps/rejected": -1.429254412651062,
|
| 675 |
+
"loss": 63.6557,
|
| 676 |
+
"nll_loss": 1.1602365970611572,
|
| 677 |
+
"rewards/accuracies": 0.715624988079071,
|
| 678 |
+
"rewards/chosen": -0.08860549330711365,
|
| 679 |
+
"rewards/margins": 0.054319947957992554,
|
| 680 |
+
"rewards/rejected": -0.1429254561662674,
|
| 681 |
+
"step": 210
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.0385185185185186,
|
| 685 |
+
"grad_norm": 119.5,
|
| 686 |
+
"learning_rate": 1.3885777771950348e-05,
|
| 687 |
+
"log_odds_chosen": 1.1024898290634155,
|
| 688 |
+
"log_odds_ratio": -0.42296814918518066,
|
| 689 |
+
"logps/chosen": -0.6938937306404114,
|
| 690 |
+
"logps/rejected": -1.399047613143921,
|
| 691 |
+
"loss": 39.1535,
|
| 692 |
+
"nll_loss": 0.9321538805961609,
|
| 693 |
+
"rewards/accuracies": 0.8125,
|
| 694 |
+
"rewards/chosen": -0.06938938051462173,
|
| 695 |
+
"rewards/margins": 0.07051538676023483,
|
| 696 |
+
"rewards/rejected": -0.13990476727485657,
|
| 697 |
+
"step": 215
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.0859259259259257,
|
| 701 |
+
"grad_norm": 151.0,
|
| 702 |
+
"learning_rate": 1.2660570475395683e-05,
|
| 703 |
+
"log_odds_chosen": 1.3341829776763916,
|
| 704 |
+
"log_odds_ratio": -0.4092956483364105,
|
| 705 |
+
"logps/chosen": -0.7508294582366943,
|
| 706 |
+
"logps/rejected": -1.6502450704574585,
|
| 707 |
+
"loss": 42.6512,
|
| 708 |
+
"nll_loss": 0.9648336172103882,
|
| 709 |
+
"rewards/accuracies": 0.824999988079071,
|
| 710 |
+
"rewards/chosen": -0.07508295029401779,
|
| 711 |
+
"rewards/margins": 0.08994156122207642,
|
| 712 |
+
"rewards/rejected": -0.16502448916435242,
|
| 713 |
+
"step": 220
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.1333333333333333,
|
| 717 |
+
"grad_norm": 129.0,
|
| 718 |
+
"learning_rate": 1.1473369030008974e-05,
|
| 719 |
+
"log_odds_chosen": 1.6117992401123047,
|
| 720 |
+
"log_odds_ratio": -0.370195209980011,
|
| 721 |
+
"logps/chosen": -0.7199097871780396,
|
| 722 |
+
"logps/rejected": -1.8543732166290283,
|
| 723 |
+
"loss": 38.9672,
|
| 724 |
+
"nll_loss": 0.9383153915405273,
|
| 725 |
+
"rewards/accuracies": 0.840624988079071,
|
| 726 |
+
"rewards/chosen": -0.07199098914861679,
|
| 727 |
+
"rewards/margins": 0.11344633251428604,
|
| 728 |
+
"rewards/rejected": -0.18543733656406403,
|
| 729 |
+
"step": 225
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.180740740740741,
|
| 733 |
+
"grad_norm": 191.0,
|
| 734 |
+
"learning_rate": 1.0327830055518842e-05,
|
| 735 |
+
"log_odds_chosen": 1.3568776845932007,
|
| 736 |
+
"log_odds_ratio": -0.373151570558548,
|
| 737 |
+
"logps/chosen": -0.7086887359619141,
|
| 738 |
+
"logps/rejected": -1.5901761054992676,
|
| 739 |
+
"loss": 40.2537,
|
| 740 |
+
"nll_loss": 0.9355926513671875,
|
| 741 |
+
"rewards/accuracies": 0.824999988079071,
|
| 742 |
+
"rewards/chosen": -0.07086886465549469,
|
| 743 |
+
"rewards/margins": 0.08814872801303864,
|
| 744 |
+
"rewards/rejected": -0.15901760756969452,
|
| 745 |
+
"step": 230
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.228148148148148,
|
| 749 |
+
"grad_norm": 208.0,
|
| 750 |
+
"learning_rate": 9.227481849865235e-06,
|
| 751 |
+
"log_odds_chosen": 1.7992521524429321,
|
| 752 |
+
"log_odds_ratio": -0.35374900698661804,
|
| 753 |
+
"logps/chosen": -0.6790289282798767,
|
| 754 |
+
"logps/rejected": -1.9464753866195679,
|
| 755 |
+
"loss": 38.095,
|
| 756 |
+
"nll_loss": 0.9396623373031616,
|
| 757 |
+
"rewards/accuracies": 0.878125011920929,
|
| 758 |
+
"rewards/chosen": -0.06790290027856827,
|
| 759 |
+
"rewards/margins": 0.12674465775489807,
|
| 760 |
+
"rewards/rejected": -0.19464752078056335,
|
| 761 |
+
"step": 235
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 2.2755555555555556,
|
| 765 |
+
"grad_norm": 616.0,
|
| 766 |
+
"learning_rate": 8.175713521924978e-06,
|
| 767 |
+
"log_odds_chosen": 1.9979493618011475,
|
| 768 |
+
"log_odds_ratio": -0.381753146648407,
|
| 769 |
+
"logps/chosen": -0.9090923070907593,
|
| 770 |
+
"logps/rejected": -2.4085757732391357,
|
| 771 |
+
"loss": 36.3664,
|
| 772 |
+
"nll_loss": 0.9206424951553345,
|
| 773 |
+
"rewards/accuracies": 0.871874988079071,
|
| 774 |
+
"rewards/chosen": -0.09090922772884369,
|
| 775 |
+
"rewards/margins": 0.1499483585357666,
|
| 776 |
+
"rewards/rejected": -0.2408575713634491,
|
| 777 |
+
"step": 240
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 2.322962962962963,
|
| 781 |
+
"grad_norm": 233.0,
|
| 782 |
+
"learning_rate": 7.1757645529443665e-06,
|
| 783 |
+
"log_odds_chosen": 2.353531837463379,
|
| 784 |
+
"log_odds_ratio": -0.3253650367259979,
|
| 785 |
+
"logps/chosen": -0.681083083152771,
|
| 786 |
+
"logps/rejected": -2.5145788192749023,
|
| 787 |
+
"loss": 35.8931,
|
| 788 |
+
"nll_loss": 0.9288511276245117,
|
| 789 |
+
"rewards/accuracies": 0.878125011920929,
|
| 790 |
+
"rewards/chosen": -0.06810830533504486,
|
| 791 |
+
"rewards/margins": 0.1833495795726776,
|
| 792 |
+
"rewards/rejected": -0.25145789980888367,
|
| 793 |
+
"step": 245
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 2.3703703703703702,
|
| 797 |
+
"grad_norm": 1592.0,
|
| 798 |
+
"learning_rate": 6.230714818829733e-06,
|
| 799 |
+
"log_odds_chosen": 2.2513911724090576,
|
| 800 |
+
"log_odds_ratio": -0.397489070892334,
|
| 801 |
+
"logps/chosen": -0.7289605140686035,
|
| 802 |
+
"logps/rejected": -2.426892042160034,
|
| 803 |
+
"loss": 38.3537,
|
| 804 |
+
"nll_loss": 0.9416518211364746,
|
| 805 |
+
"rewards/accuracies": 0.878125011920929,
|
| 806 |
+
"rewards/chosen": -0.07289605587720871,
|
| 807 |
+
"rewards/margins": 0.16979314386844635,
|
| 808 |
+
"rewards/rejected": -0.24268922209739685,
|
| 809 |
+
"step": 250
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.417777777777778,
|
| 813 |
+
"grad_norm": 476.0,
|
| 814 |
+
"learning_rate": 5.343475104027743e-06,
|
| 815 |
+
"log_odds_chosen": 3.081707000732422,
|
| 816 |
+
"log_odds_ratio": -0.616198718547821,
|
| 817 |
+
"logps/chosen": -0.9827086329460144,
|
| 818 |
+
"logps/rejected": -3.5485787391662598,
|
| 819 |
+
"loss": 36.6459,
|
| 820 |
+
"nll_loss": 0.9459680318832397,
|
| 821 |
+
"rewards/accuracies": 0.890625,
|
| 822 |
+
"rewards/chosen": -0.09827087074518204,
|
| 823 |
+
"rewards/margins": 0.25658702850341797,
|
| 824 |
+
"rewards/rejected": -0.354857861995697,
|
| 825 |
+
"step": 255
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.4651851851851854,
|
| 829 |
+
"grad_norm": 2640.0,
|
| 830 |
+
"learning_rate": 4.516778136213037e-06,
|
| 831 |
+
"log_odds_chosen": 2.5836405754089355,
|
| 832 |
+
"log_odds_ratio": -0.5884393453598022,
|
| 833 |
+
"logps/chosen": -0.9286125898361206,
|
| 834 |
+
"logps/rejected": -3.0004982948303223,
|
| 835 |
+
"loss": 36.5019,
|
| 836 |
+
"nll_loss": 0.9332065582275391,
|
| 837 |
+
"rewards/accuracies": 0.875,
|
| 838 |
+
"rewards/chosen": -0.09286126494407654,
|
| 839 |
+
"rewards/margins": 0.20718856155872345,
|
| 840 |
+
"rewards/rejected": -0.3000498414039612,
|
| 841 |
+
"step": 260
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 2.5125925925925925,
|
| 845 |
+
"grad_norm": 928.0,
|
| 846 |
+
"learning_rate": 3.7531701693965554e-06,
|
| 847 |
+
"log_odds_chosen": 2.268559694290161,
|
| 848 |
+
"log_odds_ratio": -0.5842736959457397,
|
| 849 |
+
"logps/chosen": -0.9478973150253296,
|
| 850 |
+
"logps/rejected": -2.7283573150634766,
|
| 851 |
+
"loss": 37.1403,
|
| 852 |
+
"nll_loss": 0.9361186027526855,
|
| 853 |
+
"rewards/accuracies": 0.859375,
|
| 854 |
+
"rewards/chosen": -0.09478974342346191,
|
| 855 |
+
"rewards/margins": 0.17804603278636932,
|
| 856 |
+
"rewards/rejected": -0.27283576130867004,
|
| 857 |
+
"step": 265
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 2.56,
|
| 861 |
+
"grad_norm": 326.0,
|
| 862 |
+
"learning_rate": 3.055003141378948e-06,
|
| 863 |
+
"log_odds_chosen": 2.6079001426696777,
|
| 864 |
+
"log_odds_ratio": -0.34198158979415894,
|
| 865 |
+
"logps/chosen": -0.6971701383590698,
|
| 866 |
+
"logps/rejected": -2.7689833641052246,
|
| 867 |
+
"loss": 33.3444,
|
| 868 |
+
"nll_loss": 0.9110336303710938,
|
| 869 |
+
"rewards/accuracies": 0.903124988079071,
|
| 870 |
+
"rewards/chosen": -0.06971701979637146,
|
| 871 |
+
"rewards/margins": 0.20718133449554443,
|
| 872 |
+
"rewards/rejected": -0.2768983244895935,
|
| 873 |
+
"step": 270
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 2.6074074074074076,
|
| 877 |
+
"grad_norm": 864.0,
|
| 878 |
+
"learning_rate": 2.424427429704365e-06,
|
| 879 |
+
"log_odds_chosen": 2.700164318084717,
|
| 880 |
+
"log_odds_ratio": -0.43103471398353577,
|
| 881 |
+
"logps/chosen": -0.814444899559021,
|
| 882 |
+
"logps/rejected": -3.0468616485595703,
|
| 883 |
+
"loss": 34.2692,
|
| 884 |
+
"nll_loss": 0.9250169992446899,
|
| 885 |
+
"rewards/accuracies": 0.8999999761581421,
|
| 886 |
+
"rewards/chosen": -0.08144449442625046,
|
| 887 |
+
"rewards/margins": 0.2232416868209839,
|
| 888 |
+
"rewards/rejected": -0.30468615889549255,
|
| 889 |
+
"step": 275
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 2.6548148148148147,
|
| 893 |
+
"grad_norm": 167.0,
|
| 894 |
+
"learning_rate": 1.8633852284264508e-06,
|
| 895 |
+
"log_odds_chosen": 2.282158374786377,
|
| 896 |
+
"log_odds_ratio": -0.35621577501296997,
|
| 897 |
+
"logps/chosen": -0.6996549963951111,
|
| 898 |
+
"logps/rejected": -2.4926743507385254,
|
| 899 |
+
"loss": 36.2288,
|
| 900 |
+
"nll_loss": 0.9279597997665405,
|
| 901 |
+
"rewards/accuracies": 0.859375,
|
| 902 |
+
"rewards/chosen": -0.06996549665927887,
|
| 903 |
+
"rewards/margins": 0.17930194735527039,
|
| 904 |
+
"rewards/rejected": -0.24926745891571045,
|
| 905 |
+
"step": 280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 2.7022222222222223,
|
| 909 |
+
"grad_norm": 252.0,
|
| 910 |
+
"learning_rate": 1.3736045660864034e-06,
|
| 911 |
+
"log_odds_chosen": 2.3198978900909424,
|
| 912 |
+
"log_odds_ratio": -0.4223317503929138,
|
| 913 |
+
"logps/chosen": -0.7921027541160583,
|
| 914 |
+
"logps/rejected": -2.6456892490386963,
|
| 915 |
+
"loss": 40.3585,
|
| 916 |
+
"nll_loss": 0.9682122468948364,
|
| 917 |
+
"rewards/accuracies": 0.8531249761581421,
|
| 918 |
+
"rewards/chosen": -0.07921028137207031,
|
| 919 |
+
"rewards/margins": 0.18535864353179932,
|
| 920 |
+
"rewards/rejected": -0.26456892490386963,
|
| 921 |
+
"step": 285
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.74962962962963,
|
| 925 |
+
"grad_norm": 660.0,
|
| 926 |
+
"learning_rate": 9.565939833279192e-07,
|
| 927 |
+
"log_odds_chosen": 3.0985660552978516,
|
| 928 |
+
"log_odds_ratio": -0.3719687759876251,
|
| 929 |
+
"logps/chosen": -0.8049660921096802,
|
| 930 |
+
"logps/rejected": -3.4025039672851562,
|
| 931 |
+
"loss": 39.5682,
|
| 932 |
+
"nll_loss": 0.9765946269035339,
|
| 933 |
+
"rewards/accuracies": 0.862500011920929,
|
| 934 |
+
"rewards/chosen": -0.08049660921096802,
|
| 935 |
+
"rewards/margins": 0.2597537934780121,
|
| 936 |
+
"rewards/rejected": -0.3402504026889801,
|
| 937 |
+
"step": 290
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 2.797037037037037,
|
| 941 |
+
"grad_norm": 596.0,
|
| 942 |
+
"learning_rate": 6.136378865420872e-07,
|
| 943 |
+
"log_odds_chosen": 2.389584541320801,
|
| 944 |
+
"log_odds_ratio": -0.3822096884250641,
|
| 945 |
+
"logps/chosen": -0.7572653889656067,
|
| 946 |
+
"logps/rejected": -2.6863067150115967,
|
| 947 |
+
"loss": 38.5353,
|
| 948 |
+
"nll_loss": 0.9491559267044067,
|
| 949 |
+
"rewards/accuracies": 0.840624988079071,
|
| 950 |
+
"rewards/chosen": -0.07572654634714127,
|
| 951 |
+
"rewards/margins": 0.19290415942668915,
|
| 952 |
+
"rewards/rejected": -0.2686306834220886,
|
| 953 |
+
"step": 295
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 2.8444444444444446,
|
| 957 |
+
"grad_norm": 1264.0,
|
| 958 |
+
"learning_rate": 3.45792591853214e-07,
|
| 959 |
+
"log_odds_chosen": 3.3418126106262207,
|
| 960 |
+
"log_odds_ratio": -0.4350895285606384,
|
| 961 |
+
"logps/chosen": -0.7672649025917053,
|
| 962 |
+
"logps/rejected": -3.545822858810425,
|
| 963 |
+
"loss": 34.2986,
|
| 964 |
+
"nll_loss": 0.9356390833854675,
|
| 965 |
+
"rewards/accuracies": 0.871874988079071,
|
| 966 |
+
"rewards/chosen": -0.07672648876905441,
|
| 967 |
+
"rewards/margins": 0.27785584330558777,
|
| 968 |
+
"rewards/rejected": -0.3545823097229004,
|
| 969 |
+
"step": 300
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 2.891851851851852,
|
| 973 |
+
"grad_norm": 143.0,
|
| 974 |
+
"learning_rate": 1.538830716302092e-07,
|
| 975 |
+
"log_odds_chosen": 3.44517183303833,
|
| 976 |
+
"log_odds_ratio": -0.3483870029449463,
|
| 977 |
+
"logps/chosen": -0.704784095287323,
|
| 978 |
+
"logps/rejected": -3.572923183441162,
|
| 979 |
+
"loss": 30.5972,
|
| 980 |
+
"nll_loss": 0.9035211801528931,
|
| 981 |
+
"rewards/accuracies": 0.9156249761581421,
|
| 982 |
+
"rewards/chosen": -0.0704784095287323,
|
| 983 |
+
"rewards/margins": 0.2868139147758484,
|
| 984 |
+
"rewards/rejected": -0.3572922945022583,
|
| 985 |
+
"step": 305
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 2.9392592592592592,
|
| 989 |
+
"grad_norm": 268.0,
|
| 990 |
+
"learning_rate": 3.8500413544415025e-08,
|
| 991 |
+
"log_odds_chosen": 2.904447317123413,
|
| 992 |
+
"log_odds_ratio": -0.30277663469314575,
|
| 993 |
+
"logps/chosen": -0.6428898572921753,
|
| 994 |
+
"logps/rejected": -2.9733409881591797,
|
| 995 |
+
"loss": 34.3165,
|
| 996 |
+
"nll_loss": 0.9161517024040222,
|
| 997 |
+
"rewards/accuracies": 0.90625,
|
| 998 |
+
"rewards/chosen": -0.06428898870944977,
|
| 999 |
+
"rewards/margins": 0.23304510116577148,
|
| 1000 |
+
"rewards/rejected": -0.29733413457870483,
|
| 1001 |
+
"step": 310
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 2.986666666666667,
|
| 1005 |
+
"grad_norm": 400.0,
|
| 1006 |
+
"learning_rate": 0.0,
|
| 1007 |
+
"log_odds_chosen": 2.1221396923065186,
|
| 1008 |
+
"log_odds_ratio": -0.44431638717651367,
|
| 1009 |
+
"logps/chosen": -0.7690244317054749,
|
| 1010 |
+
"logps/rejected": -2.3752024173736572,
|
| 1011 |
+
"loss": 35.9892,
|
| 1012 |
+
"nll_loss": 0.9088751077651978,
|
| 1013 |
+
"rewards/accuracies": 0.8687499761581421,
|
| 1014 |
+
"rewards/chosen": -0.07690244913101196,
|
| 1015 |
+
"rewards/margins": 0.16061779856681824,
|
| 1016 |
+
"rewards/rejected": -0.2375202178955078,
|
| 1017 |
+
"step": 315
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 2.986666666666667,
|
| 1021 |
+
"step": 315,
|
| 1022 |
+
"total_flos": 0.0,
|
| 1023 |
+
"train_loss": 100.64729357522631,
|
| 1024 |
+
"train_runtime": 7331.4996,
|
| 1025 |
+
"train_samples_per_second": 2.762,
|
| 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 |
+
}
|