Model save
Browse files- README.md +67 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +778 -0
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: google/gemma-7b
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: gemma-7b-sft-noisy-5e-5-005-v140
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- orpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for gemma-7b-sft-noisy-5e-5-005-v140
|
| 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-sft-noisy-5e-5-005-v140", 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/okcuiyak)
|
| 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: 0.15.2
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.5.1
|
| 40 |
+
- Datasets: 3.2.0
|
| 41 |
+
- Tokenizers: 0.21.1
|
| 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": 2.9881235154394297,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 28.92138784359663,
|
| 5 |
+
"train_runtime": 2488.1916,
|
| 6 |
+
"train_samples": 5050,
|
| 7 |
+
"train_samples_per_second": 6.089,
|
| 8 |
+
"train_steps_per_second": 0.094
|
| 9 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.9881235154394297,
|
| 3 |
+
"total_flos": 0.0,
|
| 4 |
+
"train_loss": 28.92138784359663,
|
| 5 |
+
"train_runtime": 2488.1916,
|
| 6 |
+
"train_samples": 5050,
|
| 7 |
+
"train_samples_per_second": 6.089,
|
| 8 |
+
"train_steps_per_second": 0.094
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,778 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9881235154394297,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 234,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.06334125098970704,
|
| 13 |
+
"grad_norm": 302.0,
|
| 14 |
+
"learning_rate": 1.0416666666666668e-05,
|
| 15 |
+
"log_odds_chosen": 3.157041072845459,
|
| 16 |
+
"log_odds_ratio": -9.87833023071289,
|
| 17 |
+
"logps/chosen": -23.06415557861328,
|
| 18 |
+
"logps/rejected": -26.220584869384766,
|
| 19 |
+
"loss": 170.2786,
|
| 20 |
+
"nll_loss": 10.642413139343262,
|
| 21 |
+
"rewards/accuracies": 0.543749988079071,
|
| 22 |
+
"rewards/chosen": -1.1532078981399536,
|
| 23 |
+
"rewards/margins": 0.15782156586647034,
|
| 24 |
+
"rewards/rejected": -1.3110291957855225,
|
| 25 |
+
"step": 5
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.12668250197941408,
|
| 29 |
+
"grad_norm": 260.0,
|
| 30 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 31 |
+
"log_odds_chosen": 6.424581050872803,
|
| 32 |
+
"log_odds_ratio": -9.701116561889648,
|
| 33 |
+
"logps/chosen": -22.176563262939453,
|
| 34 |
+
"logps/rejected": -28.598962783813477,
|
| 35 |
+
"loss": 168.5609,
|
| 36 |
+
"nll_loss": 10.535055160522461,
|
| 37 |
+
"rewards/accuracies": 0.581250011920929,
|
| 38 |
+
"rewards/chosen": -1.1088283061981201,
|
| 39 |
+
"rewards/margins": 0.32112011313438416,
|
| 40 |
+
"rewards/rejected": -1.4299482107162476,
|
| 41 |
+
"step": 10
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.19002375296912113,
|
| 45 |
+
"grad_norm": 370.0,
|
| 46 |
+
"learning_rate": 3.125e-05,
|
| 47 |
+
"log_odds_chosen": 5.751986503601074,
|
| 48 |
+
"log_odds_ratio": -6.494493007659912,
|
| 49 |
+
"logps/chosen": -18.182918548583984,
|
| 50 |
+
"logps/rejected": -23.933364868164062,
|
| 51 |
+
"loss": 135.9466,
|
| 52 |
+
"nll_loss": 8.496664047241211,
|
| 53 |
+
"rewards/accuracies": 0.546875,
|
| 54 |
+
"rewards/chosen": -0.9091460108757019,
|
| 55 |
+
"rewards/margins": 0.28752225637435913,
|
| 56 |
+
"rewards/rejected": -1.196668267250061,
|
| 57 |
+
"step": 15
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.25336500395882816,
|
| 61 |
+
"grad_norm": 174.0,
|
| 62 |
+
"learning_rate": 4.166666666666667e-05,
|
| 63 |
+
"log_odds_chosen": 1.3518788814544678,
|
| 64 |
+
"log_odds_ratio": -3.2631962299346924,
|
| 65 |
+
"logps/chosen": -8.165756225585938,
|
| 66 |
+
"logps/rejected": -9.506813049316406,
|
| 67 |
+
"loss": 67.1286,
|
| 68 |
+
"nll_loss": 4.195539951324463,
|
| 69 |
+
"rewards/accuracies": 0.581250011920929,
|
| 70 |
+
"rewards/chosen": -0.40828782320022583,
|
| 71 |
+
"rewards/margins": 0.06705284118652344,
|
| 72 |
+
"rewards/rejected": -0.47534069418907166,
|
| 73 |
+
"step": 20
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.3167062549485352,
|
| 77 |
+
"grad_norm": 226.0,
|
| 78 |
+
"learning_rate": 4.999720254525684e-05,
|
| 79 |
+
"log_odds_chosen": 0.28868618607521057,
|
| 80 |
+
"log_odds_ratio": -0.8123720288276672,
|
| 81 |
+
"logps/chosen": -1.6918294429779053,
|
| 82 |
+
"logps/rejected": -1.9460710287094116,
|
| 83 |
+
"loss": 32.5673,
|
| 84 |
+
"nll_loss": 2.0354533195495605,
|
| 85 |
+
"rewards/accuracies": 0.565625011920929,
|
| 86 |
+
"rewards/chosen": -0.08459147810935974,
|
| 87 |
+
"rewards/margins": 0.012712081894278526,
|
| 88 |
+
"rewards/rejected": -0.09730355441570282,
|
| 89 |
+
"step": 25
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.38004750593824227,
|
| 93 |
+
"grad_norm": 85.5,
|
| 94 |
+
"learning_rate": 4.989935734988098e-05,
|
| 95 |
+
"log_odds_chosen": 0.13547295331954956,
|
| 96 |
+
"log_odds_ratio": -0.8609800338745117,
|
| 97 |
+
"logps/chosen": -1.710571050643921,
|
| 98 |
+
"logps/rejected": -1.8393272161483765,
|
| 99 |
+
"loss": 31.734,
|
| 100 |
+
"nll_loss": 1.9833742380142212,
|
| 101 |
+
"rewards/accuracies": 0.546875,
|
| 102 |
+
"rewards/chosen": -0.08552855998277664,
|
| 103 |
+
"rewards/margins": 0.0064378054812550545,
|
| 104 |
+
"rewards/rejected": -0.09196637570858002,
|
| 105 |
+
"step": 30
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.4433887569279493,
|
| 109 |
+
"grad_norm": 604.0,
|
| 110 |
+
"learning_rate": 4.9662264848088034e-05,
|
| 111 |
+
"log_odds_chosen": 0.038943640887737274,
|
| 112 |
+
"log_odds_ratio": -0.864970862865448,
|
| 113 |
+
"logps/chosen": -1.499442219734192,
|
| 114 |
+
"logps/rejected": -1.5386669635772705,
|
| 115 |
+
"loss": 29.0551,
|
| 116 |
+
"nll_loss": 1.815943956375122,
|
| 117 |
+
"rewards/accuracies": 0.48750001192092896,
|
| 118 |
+
"rewards/chosen": -0.07497210055589676,
|
| 119 |
+
"rewards/margins": 0.0019612452015280724,
|
| 120 |
+
"rewards/rejected": -0.076933354139328,
|
| 121 |
+
"step": 35
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.5067300079176563,
|
| 125 |
+
"grad_norm": 69.5,
|
| 126 |
+
"learning_rate": 4.928725095732169e-05,
|
| 127 |
+
"log_odds_chosen": 0.14633053541183472,
|
| 128 |
+
"log_odds_ratio": -0.7850778102874756,
|
| 129 |
+
"logps/chosen": -1.3732013702392578,
|
| 130 |
+
"logps/rejected": -1.5010284185409546,
|
| 131 |
+
"loss": 27.5485,
|
| 132 |
+
"nll_loss": 1.7217798233032227,
|
| 133 |
+
"rewards/accuracies": 0.550000011920929,
|
| 134 |
+
"rewards/chosen": -0.06866006553173065,
|
| 135 |
+
"rewards/margins": 0.006391354836523533,
|
| 136 |
+
"rewards/rejected": -0.07505142688751221,
|
| 137 |
+
"step": 40
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.5700712589073634,
|
| 141 |
+
"grad_norm": 42.25,
|
| 142 |
+
"learning_rate": 4.877641290737884e-05,
|
| 143 |
+
"log_odds_chosen": 0.1304132044315338,
|
| 144 |
+
"log_odds_ratio": -0.7902301549911499,
|
| 145 |
+
"logps/chosen": -1.3314447402954102,
|
| 146 |
+
"logps/rejected": -1.4372903108596802,
|
| 147 |
+
"loss": 26.0469,
|
| 148 |
+
"nll_loss": 1.6279296875,
|
| 149 |
+
"rewards/accuracies": 0.534375011920929,
|
| 150 |
+
"rewards/chosen": -0.06657223403453827,
|
| 151 |
+
"rewards/margins": 0.0052922810427844524,
|
| 152 |
+
"rewards/rejected": -0.07186451554298401,
|
| 153 |
+
"step": 45
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.6334125098970704,
|
| 157 |
+
"grad_norm": 31.0,
|
| 158 |
+
"learning_rate": 4.813260751184992e-05,
|
| 159 |
+
"log_odds_chosen": 0.274090439081192,
|
| 160 |
+
"log_odds_ratio": -0.7146386504173279,
|
| 161 |
+
"logps/chosen": -1.2023499011993408,
|
| 162 |
+
"logps/rejected": -1.4235477447509766,
|
| 163 |
+
"loss": 24.6248,
|
| 164 |
+
"nll_loss": 1.539048433303833,
|
| 165 |
+
"rewards/accuracies": 0.59375,
|
| 166 |
+
"rewards/chosen": -0.06011749058961868,
|
| 167 |
+
"rewards/margins": 0.011059902608394623,
|
| 168 |
+
"rewards/rejected": -0.0711773931980133,
|
| 169 |
+
"step": 50
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.6967537608867775,
|
| 173 |
+
"grad_norm": 50.75,
|
| 174 |
+
"learning_rate": 4.7359435191658425e-05,
|
| 175 |
+
"log_odds_chosen": 0.10976365953683853,
|
| 176 |
+
"log_odds_ratio": -0.7592318058013916,
|
| 177 |
+
"logps/chosen": -1.2341073751449585,
|
| 178 |
+
"logps/rejected": -1.3237587213516235,
|
| 179 |
+
"loss": 24.4883,
|
| 180 |
+
"nll_loss": 1.5305156707763672,
|
| 181 |
+
"rewards/accuracies": 0.5093749761581421,
|
| 182 |
+
"rewards/chosen": -0.06170536205172539,
|
| 183 |
+
"rewards/margins": 0.004482570104300976,
|
| 184 |
+
"rewards/rejected": -0.06618793308734894,
|
| 185 |
+
"step": 55
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.7600950118764845,
|
| 189 |
+
"grad_norm": 73.5,
|
| 190 |
+
"learning_rate": 4.6461219840046654e-05,
|
| 191 |
+
"log_odds_chosen": 0.19487467408180237,
|
| 192 |
+
"log_odds_ratio": -0.701287567615509,
|
| 193 |
+
"logps/chosen": -1.1585313081741333,
|
| 194 |
+
"logps/rejected": -1.3133569955825806,
|
| 195 |
+
"loss": 24.4529,
|
| 196 |
+
"nll_loss": 1.5283066034317017,
|
| 197 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 198 |
+
"rewards/chosen": -0.05792655795812607,
|
| 199 |
+
"rewards/margins": 0.007741288281977177,
|
| 200 |
+
"rewards/rejected": -0.06566785275936127,
|
| 201 |
+
"step": 60
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.8234362628661916,
|
| 205 |
+
"grad_norm": 52.25,
|
| 206 |
+
"learning_rate": 4.544298464161079e-05,
|
| 207 |
+
"log_odds_chosen": 0.24091807007789612,
|
| 208 |
+
"log_odds_ratio": -0.6930054426193237,
|
| 209 |
+
"logps/chosen": -1.125671148300171,
|
| 210 |
+
"logps/rejected": -1.3065881729125977,
|
| 211 |
+
"loss": 24.4496,
|
| 212 |
+
"nll_loss": 1.5281010866165161,
|
| 213 |
+
"rewards/accuracies": 0.578125,
|
| 214 |
+
"rewards/chosen": -0.05628355219960213,
|
| 215 |
+
"rewards/margins": 0.009045850485563278,
|
| 216 |
+
"rewards/rejected": -0.06532939523458481,
|
| 217 |
+
"step": 65
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 0.8867775138558986,
|
| 221 |
+
"grad_norm": 43.75,
|
| 222 |
+
"learning_rate": 4.431042398061499e-05,
|
| 223 |
+
"log_odds_chosen": 0.09587414562702179,
|
| 224 |
+
"log_odds_ratio": -0.7694802284240723,
|
| 225 |
+
"logps/chosen": -1.169614553451538,
|
| 226 |
+
"logps/rejected": -1.2482807636260986,
|
| 227 |
+
"loss": 24.5127,
|
| 228 |
+
"nll_loss": 1.5320452451705933,
|
| 229 |
+
"rewards/accuracies": 0.5218750238418579,
|
| 230 |
+
"rewards/chosen": -0.058480728417634964,
|
| 231 |
+
"rewards/margins": 0.003933308180421591,
|
| 232 |
+
"rewards/rejected": -0.06241403892636299,
|
| 233 |
+
"step": 70
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.9501187648456056,
|
| 237 |
+
"grad_norm": 38.5,
|
| 238 |
+
"learning_rate": 4.306987159568479e-05,
|
| 239 |
+
"log_odds_chosen": 0.16929097473621368,
|
| 240 |
+
"log_odds_ratio": -0.7172764539718628,
|
| 241 |
+
"logps/chosen": -1.0788049697875977,
|
| 242 |
+
"logps/rejected": -1.1940680742263794,
|
| 243 |
+
"loss": 22.7263,
|
| 244 |
+
"nll_loss": 1.4203910827636719,
|
| 245 |
+
"rewards/accuracies": 0.59375,
|
| 246 |
+
"rewards/chosen": -0.05394024774432182,
|
| 247 |
+
"rewards/margins": 0.005763149820268154,
|
| 248 |
+
"rewards/rejected": -0.05970339849591255,
|
| 249 |
+
"step": 75
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.0253365003958828,
|
| 253 |
+
"grad_norm": 181.0,
|
| 254 |
+
"learning_rate": 4.172826515897146e-05,
|
| 255 |
+
"log_odds_chosen": 0.32294443249702454,
|
| 256 |
+
"log_odds_ratio": -0.6628458499908447,
|
| 257 |
+
"logps/chosen": -1.0215741395950317,
|
| 258 |
+
"logps/rejected": -1.248105525970459,
|
| 259 |
+
"loss": 25.0471,
|
| 260 |
+
"nll_loss": 1.3915075063705444,
|
| 261 |
+
"rewards/accuracies": 0.605555534362793,
|
| 262 |
+
"rewards/chosen": -0.05107870697975159,
|
| 263 |
+
"rewards/margins": 0.011326568201184273,
|
| 264 |
+
"rewards/rejected": -0.06240528076887131,
|
| 265 |
+
"step": 80
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 1.0886777513855899,
|
| 269 |
+
"grad_norm": 24.625,
|
| 270 |
+
"learning_rate": 4.029310747787516e-05,
|
| 271 |
+
"log_odds_chosen": 0.540235698223114,
|
| 272 |
+
"log_odds_ratio": -0.5549238324165344,
|
| 273 |
+
"logps/chosen": -0.8654016256332397,
|
| 274 |
+
"logps/rejected": -1.20816969871521,
|
| 275 |
+
"loss": 19.002,
|
| 276 |
+
"nll_loss": 1.1876252889633179,
|
| 277 |
+
"rewards/accuracies": 0.7093750238418579,
|
| 278 |
+
"rewards/chosen": -0.043270088732242584,
|
| 279 |
+
"rewards/margins": 0.01713840290904045,
|
| 280 |
+
"rewards/rejected": -0.06040848419070244,
|
| 281 |
+
"step": 85
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.152019002375297,
|
| 285 |
+
"grad_norm": 31.5,
|
| 286 |
+
"learning_rate": 3.8772424536302564e-05,
|
| 287 |
+
"log_odds_chosen": 0.6173906326293945,
|
| 288 |
+
"log_odds_ratio": -0.5311087965965271,
|
| 289 |
+
"logps/chosen": -0.7949373722076416,
|
| 290 |
+
"logps/rejected": -1.1982877254486084,
|
| 291 |
+
"loss": 18.4072,
|
| 292 |
+
"nll_loss": 1.1504485607147217,
|
| 293 |
+
"rewards/accuracies": 0.7281249761581421,
|
| 294 |
+
"rewards/chosen": -0.03974687308073044,
|
| 295 |
+
"rewards/margins": 0.02016752026975155,
|
| 296 |
+
"rewards/rejected": -0.05991438776254654,
|
| 297 |
+
"step": 90
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.2153602533650039,
|
| 301 |
+
"grad_norm": 58.5,
|
| 302 |
+
"learning_rate": 3.717472061010918e-05,
|
| 303 |
+
"log_odds_chosen": 0.5371578931808472,
|
| 304 |
+
"log_odds_ratio": -0.5785112380981445,
|
| 305 |
+
"logps/chosen": -0.8111588358879089,
|
| 306 |
+
"logps/rejected": -1.1321982145309448,
|
| 307 |
+
"loss": 18.4027,
|
| 308 |
+
"nll_loss": 1.1501662731170654,
|
| 309 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 310 |
+
"rewards/chosen": -0.040557943284511566,
|
| 311 |
+
"rewards/margins": 0.016051968559622765,
|
| 312 |
+
"rewards/rejected": -0.05660990998148918,
|
| 313 |
+
"step": 95
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 1.278701504354711,
|
| 317 |
+
"grad_norm": 43.25,
|
| 318 |
+
"learning_rate": 3.550893070773914e-05,
|
| 319 |
+
"log_odds_chosen": 0.5756627321243286,
|
| 320 |
+
"log_odds_ratio": -0.5552070736885071,
|
| 321 |
+
"logps/chosen": -0.8670624494552612,
|
| 322 |
+
"logps/rejected": -1.227243423461914,
|
| 323 |
+
"loss": 18.8739,
|
| 324 |
+
"nll_loss": 1.1796166896820068,
|
| 325 |
+
"rewards/accuracies": 0.690625011920929,
|
| 326 |
+
"rewards/chosen": -0.0433531180024147,
|
| 327 |
+
"rewards/margins": 0.01800905168056488,
|
| 328 |
+
"rewards/rejected": -0.061362169682979584,
|
| 329 |
+
"step": 100
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.342042755344418,
|
| 333 |
+
"grad_norm": 24.25,
|
| 334 |
+
"learning_rate": 3.378437060203357e-05,
|
| 335 |
+
"log_odds_chosen": 0.4615544378757477,
|
| 336 |
+
"log_odds_ratio": -0.5932336449623108,
|
| 337 |
+
"logps/chosen": -0.8293741941452026,
|
| 338 |
+
"logps/rejected": -1.1050505638122559,
|
| 339 |
+
"loss": 18.096,
|
| 340 |
+
"nll_loss": 1.1310014724731445,
|
| 341 |
+
"rewards/accuracies": 0.671875,
|
| 342 |
+
"rewards/chosen": -0.04146871715784073,
|
| 343 |
+
"rewards/margins": 0.01378381997346878,
|
| 344 |
+
"rewards/rejected": -0.05525252968072891,
|
| 345 |
+
"step": 105
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.405384006334125,
|
| 349 |
+
"grad_norm": 54.5,
|
| 350 |
+
"learning_rate": 3.201068473265007e-05,
|
| 351 |
+
"log_odds_chosen": 0.49423256516456604,
|
| 352 |
+
"log_odds_ratio": -0.5842307209968567,
|
| 353 |
+
"logps/chosen": -0.8439785242080688,
|
| 354 |
+
"logps/rejected": -1.1538585424423218,
|
| 355 |
+
"loss": 19.262,
|
| 356 |
+
"nll_loss": 1.2038729190826416,
|
| 357 |
+
"rewards/accuracies": 0.671875,
|
| 358 |
+
"rewards/chosen": -0.04219892621040344,
|
| 359 |
+
"rewards/margins": 0.015493998304009438,
|
| 360 |
+
"rewards/rejected": -0.05769292637705803,
|
| 361 |
+
"step": 110
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.468725257323832,
|
| 365 |
+
"grad_norm": 28.5,
|
| 366 |
+
"learning_rate": 3.0197792270443982e-05,
|
| 367 |
+
"log_odds_chosen": 0.5744311809539795,
|
| 368 |
+
"log_odds_ratio": -0.5467215776443481,
|
| 369 |
+
"logps/chosen": -0.7895219326019287,
|
| 370 |
+
"logps/rejected": -1.148335337638855,
|
| 371 |
+
"loss": 18.1877,
|
| 372 |
+
"nll_loss": 1.1367297172546387,
|
| 373 |
+
"rewards/accuracies": 0.731249988079071,
|
| 374 |
+
"rewards/chosen": -0.039476100355386734,
|
| 375 |
+
"rewards/margins": 0.017940666526556015,
|
| 376 |
+
"rewards/rejected": -0.05741675943136215,
|
| 377 |
+
"step": 115
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.5320665083135392,
|
| 381 |
+
"grad_norm": 56.75,
|
| 382 |
+
"learning_rate": 2.8355831645441388e-05,
|
| 383 |
+
"log_odds_chosen": 0.6337670683860779,
|
| 384 |
+
"log_odds_ratio": -0.5344417691230774,
|
| 385 |
+
"logps/chosen": -0.7838994860649109,
|
| 386 |
+
"logps/rejected": -1.1551088094711304,
|
| 387 |
+
"loss": 18.0842,
|
| 388 |
+
"nll_loss": 1.1302621364593506,
|
| 389 |
+
"rewards/accuracies": 0.6968749761581421,
|
| 390 |
+
"rewards/chosen": -0.0391949787735939,
|
| 391 |
+
"rewards/margins": 0.018560461699962616,
|
| 392 |
+
"rewards/rejected": -0.05775544047355652,
|
| 393 |
+
"step": 120
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 1.5954077593032463,
|
| 397 |
+
"grad_norm": 41.75,
|
| 398 |
+
"learning_rate": 2.649510384862586e-05,
|
| 399 |
+
"log_odds_chosen": 0.5285617709159851,
|
| 400 |
+
"log_odds_ratio": -0.5650445818901062,
|
| 401 |
+
"logps/chosen": -0.7971245646476746,
|
| 402 |
+
"logps/rejected": -1.1233524084091187,
|
| 403 |
+
"loss": 18.0659,
|
| 404 |
+
"nll_loss": 1.12911856174469,
|
| 405 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 406 |
+
"rewards/chosen": -0.03985622525215149,
|
| 407 |
+
"rewards/margins": 0.016311392188072205,
|
| 408 |
+
"rewards/rejected": -0.05616762116551399,
|
| 409 |
+
"step": 125
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.6587490102929534,
|
| 413 |
+
"grad_norm": 52.25,
|
| 414 |
+
"learning_rate": 2.4626014824618415e-05,
|
| 415 |
+
"log_odds_chosen": 0.4448211193084717,
|
| 416 |
+
"log_odds_ratio": -0.5954711437225342,
|
| 417 |
+
"logps/chosen": -0.7822165489196777,
|
| 418 |
+
"logps/rejected": -1.0468573570251465,
|
| 419 |
+
"loss": 18.0461,
|
| 420 |
+
"nll_loss": 1.1278802156448364,
|
| 421 |
+
"rewards/accuracies": 0.653124988079071,
|
| 422 |
+
"rewards/chosen": -0.039110831916332245,
|
| 423 |
+
"rewards/margins": 0.013232040219008923,
|
| 424 |
+
"rewards/rejected": -0.052342869341373444,
|
| 425 |
+
"step": 130
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.7220902612826603,
|
| 429 |
+
"grad_norm": 38.75,
|
| 430 |
+
"learning_rate": 2.2759017277414166e-05,
|
| 431 |
+
"log_odds_chosen": 0.532827615737915,
|
| 432 |
+
"log_odds_ratio": -0.5476608276367188,
|
| 433 |
+
"logps/chosen": -0.8188691139221191,
|
| 434 |
+
"logps/rejected": -1.1497671604156494,
|
| 435 |
+
"loss": 18.2174,
|
| 436 |
+
"nll_loss": 1.1385877132415771,
|
| 437 |
+
"rewards/accuracies": 0.706250011920929,
|
| 438 |
+
"rewards/chosen": -0.040943462401628494,
|
| 439 |
+
"rewards/margins": 0.016544902697205544,
|
| 440 |
+
"rewards/rejected": -0.05748836323618889,
|
| 441 |
+
"step": 135
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 1.7854315122723674,
|
| 445 |
+
"grad_norm": 38.25,
|
| 446 |
+
"learning_rate": 2.090455221462156e-05,
|
| 447 |
+
"log_odds_chosen": 0.551643967628479,
|
| 448 |
+
"log_odds_ratio": -0.5697082281112671,
|
| 449 |
+
"logps/chosen": -0.7936131358146667,
|
| 450 |
+
"logps/rejected": -1.1227937936782837,
|
| 451 |
+
"loss": 17.6124,
|
| 452 |
+
"nll_loss": 1.1007720232009888,
|
| 453 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 454 |
+
"rewards/chosen": -0.03968065604567528,
|
| 455 |
+
"rewards/margins": 0.016459036618471146,
|
| 456 |
+
"rewards/rejected": -0.05613969638943672,
|
| 457 |
+
"step": 140
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.8487727632620743,
|
| 461 |
+
"grad_norm": 21.25,
|
| 462 |
+
"learning_rate": 1.9072990557112564e-05,
|
| 463 |
+
"log_odds_chosen": 0.7169318795204163,
|
| 464 |
+
"log_odds_ratio": -0.5285918712615967,
|
| 465 |
+
"logps/chosen": -0.768589198589325,
|
| 466 |
+
"logps/rejected": -1.2199491262435913,
|
| 467 |
+
"loss": 17.5301,
|
| 468 |
+
"nll_loss": 1.0956323146820068,
|
| 469 |
+
"rewards/accuracies": 0.737500011920929,
|
| 470 |
+
"rewards/chosen": -0.03842946141958237,
|
| 471 |
+
"rewards/margins": 0.022568000480532646,
|
| 472 |
+
"rewards/rejected": -0.06099746376276016,
|
| 473 |
+
"step": 145
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.9121140142517814,
|
| 477 |
+
"grad_norm": 20.625,
|
| 478 |
+
"learning_rate": 1.7274575140626318e-05,
|
| 479 |
+
"log_odds_chosen": 0.5951381921768188,
|
| 480 |
+
"log_odds_ratio": -0.5281034708023071,
|
| 481 |
+
"logps/chosen": -0.7978143692016602,
|
| 482 |
+
"logps/rejected": -1.146867036819458,
|
| 483 |
+
"loss": 17.1084,
|
| 484 |
+
"nll_loss": 1.0692765712738037,
|
| 485 |
+
"rewards/accuracies": 0.731249988079071,
|
| 486 |
+
"rewards/chosen": -0.03989071771502495,
|
| 487 |
+
"rewards/margins": 0.01745263859629631,
|
| 488 |
+
"rewards/rejected": -0.05734335258603096,
|
| 489 |
+
"step": 150
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 1.9754552652414885,
|
| 493 |
+
"grad_norm": 25.875,
|
| 494 |
+
"learning_rate": 1.551936343367679e-05,
|
| 495 |
+
"log_odds_chosen": 0.5835083723068237,
|
| 496 |
+
"log_odds_ratio": -0.5596362948417664,
|
| 497 |
+
"logps/chosen": -0.793925404548645,
|
| 498 |
+
"logps/rejected": -1.142182469367981,
|
| 499 |
+
"loss": 18.0845,
|
| 500 |
+
"nll_loss": 1.1302804946899414,
|
| 501 |
+
"rewards/accuracies": 0.690625011920929,
|
| 502 |
+
"rewards/chosen": -0.03969626873731613,
|
| 503 |
+
"rewards/margins": 0.017412852495908737,
|
| 504 |
+
"rewards/rejected": -0.05710912495851517,
|
| 505 |
+
"step": 155
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 2.0506730007917655,
|
| 509 |
+
"grad_norm": 20.5,
|
| 510 |
+
"learning_rate": 1.3817171292109183e-05,
|
| 511 |
+
"log_odds_chosen": 0.8345562219619751,
|
| 512 |
+
"log_odds_ratio": -0.46272528171539307,
|
| 513 |
+
"logps/chosen": -0.6592908501625061,
|
| 514 |
+
"logps/rejected": -1.1182148456573486,
|
| 515 |
+
"loss": 17.5074,
|
| 516 |
+
"nll_loss": 0.9726357460021973,
|
| 517 |
+
"rewards/accuracies": 0.7972221970558167,
|
| 518 |
+
"rewards/chosen": -0.032964546233415604,
|
| 519 |
+
"rewards/margins": 0.022946199402213097,
|
| 520 |
+
"rewards/rejected": -0.055910736322402954,
|
| 521 |
+
"step": 160
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 2.1140142517814726,
|
| 525 |
+
"grad_norm": 39.5,
|
| 526 |
+
"learning_rate": 1.217751806485235e-05,
|
| 527 |
+
"log_odds_chosen": 1.2292115688323975,
|
| 528 |
+
"log_odds_ratio": -0.3634593188762665,
|
| 529 |
+
"logps/chosen": -0.5913038849830627,
|
| 530 |
+
"logps/rejected": -1.2762850522994995,
|
| 531 |
+
"loss": 14.5297,
|
| 532 |
+
"nll_loss": 0.9081088304519653,
|
| 533 |
+
"rewards/accuracies": 0.862500011920929,
|
| 534 |
+
"rewards/chosen": -0.029565196484327316,
|
| 535 |
+
"rewards/margins": 0.034249067306518555,
|
| 536 |
+
"rewards/rejected": -0.06381426751613617,
|
| 537 |
+
"step": 165
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 2.1773555027711797,
|
| 541 |
+
"grad_norm": 24.375,
|
| 542 |
+
"learning_rate": 1.0609573357858166e-05,
|
| 543 |
+
"log_odds_chosen": 1.1033363342285156,
|
| 544 |
+
"log_odds_ratio": -0.3850035071372986,
|
| 545 |
+
"logps/chosen": -0.5864080786705017,
|
| 546 |
+
"logps/rejected": -1.1940311193466187,
|
| 547 |
+
"loss": 13.3906,
|
| 548 |
+
"nll_loss": 0.8369148969650269,
|
| 549 |
+
"rewards/accuracies": 0.831250011920929,
|
| 550 |
+
"rewards/chosen": -0.029320403933525085,
|
| 551 |
+
"rewards/margins": 0.030381154268980026,
|
| 552 |
+
"rewards/rejected": -0.05970155447721481,
|
| 553 |
+
"step": 170
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 2.240696753760887,
|
| 557 |
+
"grad_norm": 26.125,
|
| 558 |
+
"learning_rate": 9.122105753945531e-06,
|
| 559 |
+
"log_odds_chosen": 1.1297001838684082,
|
| 560 |
+
"log_odds_ratio": -0.3876176178455353,
|
| 561 |
+
"logps/chosen": -0.5929871797561646,
|
| 562 |
+
"logps/rejected": -1.2141138315200806,
|
| 563 |
+
"loss": 13.93,
|
| 564 |
+
"nll_loss": 0.8706231117248535,
|
| 565 |
+
"rewards/accuracies": 0.8218749761581421,
|
| 566 |
+
"rewards/chosen": -0.02964935638010502,
|
| 567 |
+
"rewards/margins": 0.03105633333325386,
|
| 568 |
+
"rewards/rejected": -0.06070569157600403,
|
| 569 |
+
"step": 175
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 2.304038004750594,
|
| 573 |
+
"grad_norm": 23.0,
|
| 574 |
+
"learning_rate": 7.723433775328384e-06,
|
| 575 |
+
"log_odds_chosen": 1.1585485935211182,
|
| 576 |
+
"log_odds_ratio": -0.38127031922340393,
|
| 577 |
+
"logps/chosen": -0.5841962695121765,
|
| 578 |
+
"logps/rejected": -1.200871467590332,
|
| 579 |
+
"loss": 13.7727,
|
| 580 |
+
"nll_loss": 0.8607965707778931,
|
| 581 |
+
"rewards/accuracies": 0.84375,
|
| 582 |
+
"rewards/chosen": -0.029209813103079796,
|
| 583 |
+
"rewards/margins": 0.030833765864372253,
|
| 584 |
+
"rewards/rejected": -0.0600435733795166,
|
| 585 |
+
"step": 180
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 2.367379255740301,
|
| 589 |
+
"grad_norm": 42.0,
|
| 590 |
+
"learning_rate": 6.421379363065142e-06,
|
| 591 |
+
"log_odds_chosen": 1.0362011194229126,
|
| 592 |
+
"log_odds_ratio": -0.39413753151893616,
|
| 593 |
+
"logps/chosen": -0.6054643988609314,
|
| 594 |
+
"logps/rejected": -1.1436563730239868,
|
| 595 |
+
"loss": 13.6624,
|
| 596 |
+
"nll_loss": 0.8538990020751953,
|
| 597 |
+
"rewards/accuracies": 0.8343750238418579,
|
| 598 |
+
"rewards/chosen": -0.03027322329580784,
|
| 599 |
+
"rewards/margins": 0.0269095990806818,
|
| 600 |
+
"rewards/rejected": -0.05718282610177994,
|
| 601 |
+
"step": 185
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 2.4307205067300077,
|
| 605 |
+
"grad_norm": 25.5,
|
| 606 |
+
"learning_rate": 5.223224133591476e-06,
|
| 607 |
+
"log_odds_chosen": 1.0900369882583618,
|
| 608 |
+
"log_odds_ratio": -0.4055331349372864,
|
| 609 |
+
"logps/chosen": -0.5883392095565796,
|
| 610 |
+
"logps/rejected": -1.1630512475967407,
|
| 611 |
+
"loss": 13.5892,
|
| 612 |
+
"nll_loss": 0.849326491355896,
|
| 613 |
+
"rewards/accuracies": 0.84375,
|
| 614 |
+
"rewards/chosen": -0.02941695973277092,
|
| 615 |
+
"rewards/margins": 0.028735607862472534,
|
| 616 |
+
"rewards/rejected": -0.058152567595243454,
|
| 617 |
+
"step": 190
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 2.494061757719715,
|
| 621 |
+
"grad_norm": 22.625,
|
| 622 |
+
"learning_rate": 4.135668656967434e-06,
|
| 623 |
+
"log_odds_chosen": 1.1677565574645996,
|
| 624 |
+
"log_odds_ratio": -0.381297767162323,
|
| 625 |
+
"logps/chosen": -0.594551682472229,
|
| 626 |
+
"logps/rejected": -1.240464687347412,
|
| 627 |
+
"loss": 13.8641,
|
| 628 |
+
"nll_loss": 0.8665077090263367,
|
| 629 |
+
"rewards/accuracies": 0.856249988079071,
|
| 630 |
+
"rewards/chosen": -0.02972758375108242,
|
| 631 |
+
"rewards/margins": 0.032295651733875275,
|
| 632 |
+
"rewards/rejected": -0.062023233622312546,
|
| 633 |
+
"step": 195
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.557403008709422,
|
| 637 |
+
"grad_norm": 24.875,
|
| 638 |
+
"learning_rate": 3.164794984571759e-06,
|
| 639 |
+
"log_odds_chosen": 1.1639173030853271,
|
| 640 |
+
"log_odds_ratio": -0.39818206429481506,
|
| 641 |
+
"logps/chosen": -0.5963728427886963,
|
| 642 |
+
"logps/rejected": -1.2124865055084229,
|
| 643 |
+
"loss": 14.1579,
|
| 644 |
+
"nll_loss": 0.8848661184310913,
|
| 645 |
+
"rewards/accuracies": 0.8187500238418579,
|
| 646 |
+
"rewards/chosen": -0.029818642884492874,
|
| 647 |
+
"rewards/margins": 0.030805686488747597,
|
| 648 |
+
"rewards/rejected": -0.06062433123588562,
|
| 649 |
+
"step": 200
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 2.620744259699129,
|
| 653 |
+
"grad_norm": 21.0,
|
| 654 |
+
"learning_rate": 2.316032635803378e-06,
|
| 655 |
+
"log_odds_chosen": 1.259031057357788,
|
| 656 |
+
"log_odds_ratio": -0.3424246907234192,
|
| 657 |
+
"logps/chosen": -0.546891450881958,
|
| 658 |
+
"logps/rejected": -1.2267786264419556,
|
| 659 |
+
"loss": 13.1764,
|
| 660 |
+
"nll_loss": 0.823525607585907,
|
| 661 |
+
"rewards/accuracies": 0.887499988079071,
|
| 662 |
+
"rewards/chosen": -0.02734457328915596,
|
| 663 |
+
"rewards/margins": 0.03399435430765152,
|
| 664 |
+
"rewards/rejected": -0.06133893132209778,
|
| 665 |
+
"step": 205
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 2.684085510688836,
|
| 669 |
+
"grad_norm": 25.625,
|
| 670 |
+
"learning_rate": 1.59412823400657e-06,
|
| 671 |
+
"log_odds_chosen": 1.2532789707183838,
|
| 672 |
+
"log_odds_ratio": -0.3643453121185303,
|
| 673 |
+
"logps/chosen": -0.5780880451202393,
|
| 674 |
+
"logps/rejected": -1.268188714981079,
|
| 675 |
+
"loss": 13.6232,
|
| 676 |
+
"nll_loss": 0.8514481782913208,
|
| 677 |
+
"rewards/accuracies": 0.8656250238418579,
|
| 678 |
+
"rewards/chosen": -0.028904402628540993,
|
| 679 |
+
"rewards/margins": 0.03450504317879677,
|
| 680 |
+
"rewards/rejected": -0.06340944021940231,
|
| 681 |
+
"step": 210
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.7474267616785433,
|
| 685 |
+
"grad_norm": 27.625,
|
| 686 |
+
"learning_rate": 1.0031189614277765e-06,
|
| 687 |
+
"log_odds_chosen": 1.122249722480774,
|
| 688 |
+
"log_odds_ratio": -0.4062274396419525,
|
| 689 |
+
"logps/chosen": -0.5947844386100769,
|
| 690 |
+
"logps/rejected": -1.1917502880096436,
|
| 691 |
+
"loss": 13.6413,
|
| 692 |
+
"nll_loss": 0.8525827527046204,
|
| 693 |
+
"rewards/accuracies": 0.8656250238418579,
|
| 694 |
+
"rewards/chosen": -0.029739221557974815,
|
| 695 |
+
"rewards/margins": 0.029848288744688034,
|
| 696 |
+
"rewards/rejected": -0.0595875158905983,
|
| 697 |
+
"step": 215
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.81076801266825,
|
| 701 |
+
"grad_norm": 27.125,
|
| 702 |
+
"learning_rate": 5.463099816548579e-07,
|
| 703 |
+
"log_odds_chosen": 1.1767040491104126,
|
| 704 |
+
"log_odds_ratio": -0.3847493529319763,
|
| 705 |
+
"logps/chosen": -0.5844542384147644,
|
| 706 |
+
"logps/rejected": -1.234933853149414,
|
| 707 |
+
"loss": 13.8583,
|
| 708 |
+
"nll_loss": 0.8661431074142456,
|
| 709 |
+
"rewards/accuracies": 0.8500000238418579,
|
| 710 |
+
"rewards/chosen": -0.02922271564602852,
|
| 711 |
+
"rewards/margins": 0.032523978501558304,
|
| 712 |
+
"rewards/rejected": -0.06174669414758682,
|
| 713 |
+
"step": 220
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.8741092636579575,
|
| 717 |
+
"grad_norm": 26.0,
|
| 718 |
+
"learning_rate": 2.262559558016325e-07,
|
| 719 |
+
"log_odds_chosen": 1.1544322967529297,
|
| 720 |
+
"log_odds_ratio": -0.37349843978881836,
|
| 721 |
+
"logps/chosen": -0.5899974703788757,
|
| 722 |
+
"logps/rejected": -1.2202789783477783,
|
| 723 |
+
"loss": 13.6775,
|
| 724 |
+
"nll_loss": 0.8548417091369629,
|
| 725 |
+
"rewards/accuracies": 0.859375,
|
| 726 |
+
"rewards/chosen": -0.029499869793653488,
|
| 727 |
+
"rewards/margins": 0.03151407092809677,
|
| 728 |
+
"rewards/rejected": -0.061013948172330856,
|
| 729 |
+
"step": 225
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.937450514647664,
|
| 733 |
+
"grad_norm": 20.375,
|
| 734 |
+
"learning_rate": 4.474675580662113e-08,
|
| 735 |
+
"log_odds_chosen": 1.106380820274353,
|
| 736 |
+
"log_odds_ratio": -0.38155874609947205,
|
| 737 |
+
"logps/chosen": -0.5873739719390869,
|
| 738 |
+
"logps/rejected": -1.173477053642273,
|
| 739 |
+
"loss": 13.7756,
|
| 740 |
+
"nll_loss": 0.8609777688980103,
|
| 741 |
+
"rewards/accuracies": 0.859375,
|
| 742 |
+
"rewards/chosen": -0.029368694871664047,
|
| 743 |
+
"rewards/margins": 0.02930515632033348,
|
| 744 |
+
"rewards/rejected": -0.05867385119199753,
|
| 745 |
+
"step": 230
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.9881235154394297,
|
| 749 |
+
"step": 234,
|
| 750 |
+
"total_flos": 0.0,
|
| 751 |
+
"train_loss": 28.92138784359663,
|
| 752 |
+
"train_runtime": 2488.1916,
|
| 753 |
+
"train_samples_per_second": 6.089,
|
| 754 |
+
"train_steps_per_second": 0.094
|
| 755 |
+
}
|
| 756 |
+
],
|
| 757 |
+
"logging_steps": 5,
|
| 758 |
+
"max_steps": 234,
|
| 759 |
+
"num_input_tokens_seen": 0,
|
| 760 |
+
"num_train_epochs": 3,
|
| 761 |
+
"save_steps": 100,
|
| 762 |
+
"stateful_callbacks": {
|
| 763 |
+
"TrainerControl": {
|
| 764 |
+
"args": {
|
| 765 |
+
"should_epoch_stop": false,
|
| 766 |
+
"should_evaluate": false,
|
| 767 |
+
"should_log": false,
|
| 768 |
+
"should_save": true,
|
| 769 |
+
"should_training_stop": true
|
| 770 |
+
},
|
| 771 |
+
"attributes": {}
|
| 772 |
+
}
|
| 773 |
+
},
|
| 774 |
+
"total_flos": 0.0,
|
| 775 |
+
"train_batch_size": 1,
|
| 776 |
+
"trial_name": null,
|
| 777 |
+
"trial_params": null
|
| 778 |
+
}
|