Downloads · 30 days
24
5% of all-time downloads
Pythia-2.8B-TLDR-Iterative-SamPO
jiazhengli/Pythia-2.8B-TLDR-Iterative-SamPO
Pythia-2.8B-TLDR-Iterative-SamPO is a text generation model from jiazhengli. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
This repository provides a fine-tuned version of Pythia-2.8B, using our proposed SamPO algorithm: Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence.
All-time downloads
442
Public
Parameters
2.8B
5.6 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors5.6 GB · 100%
At a glance
- Task
- Text Generation
- Library
- transformers
- License
- apache-2.0
- Model type
- gpt_neox
- Access
- Public
- Created
- Jun 17, 2024
- Updated
- Jun 17, 2024
- SHA
- 5abb14f9
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
Model Card for Pythia-2.8B-TLDR-Iterative-SamPO
This repository provides a fine-tuned version of Pythia-2.8B, using our proposed SamPO algorithm: Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence.
Performance
| vs. SFT | wins | len / token |
|---|---|---|
| DPO | 60.98 | 53.8 |
| Iterative DPO | 73.58 | 66.65 |
| Length Normed DPO | 58.13 | 47.34 |
| SimPO | 33.33 | 31.9 |
| Iterative SamPO | 73.58 | 49.54 |
Evaluation Details
We test our model with the same GPT-4 Win rate prompt template proposed by the DPO paper. The sampled test set is included in this repo.
Training hyperparameters
The following hyperparameters were used during DPO/SamPO training:
- DPO beta: 0.5
- learning_rate: 1e-6
- total_train_batch_size: 128
- optimizer: AdamW with beta1 0.9, beta2 0.999 and epsilon 1e-8
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- Weight Decay: 0.0
- num_epochs: 1.0
- Task
- Text Generation
- Library
- transformers
- Type
- gpt_neox
- License
- apache-2.0
- Languages
- en
- Created
- Jun 17, 2024
- Updated
- Jun 17, 2024