Downloads · 30 days
1
3% of all-time downloads
model_hh_shp2_dpo9
guoyu-zhang/model_hh_shp2_dpo9
model_hh_shp2_dpo9 is a machine learning model from guoyu-zhang. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
should probably proofread and complete it, then remove this comment. --
All-time downloads
29
Public
Repo size
12.6 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors25.2 MB · 93%
At a glance
- Library
- peft
- Access
- Public
- Created
- Apr 16, 2024
- Updated
- Apr 16, 2024
- SHA
- b9747658
Base models
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
model_hh_shp2_dpo9
This model is a fine-tuned version of meta-llama/Llama-2-7b-chat-hf on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.7428
- Rewards/chosen: -0.9061
- Rewards/rejected: -2.4348
- Rewards/accuracies: 0.5100
- Rewards/margins: 1.5287
- Logps/rejected: -234.8804
- Logps/chosen: -250.6243
- Logits/rejected: -0.7744
- Logits/chosen: -0.7754
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0005
- train_batch_size: 4
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0301 | 2.67 | 100 | 1.9636 | -1.8105 | -3.0008 | 0.5900 | 1.1904 | -235.5094 | -251.6292 | -0.6212 | -0.6500 |
| 0.0414 | 5.33 | 200 | 3.3526 | -0.0738 | -1.1593 | 0.5400 | 1.0855 | -233.4633 | -249.6996 | -0.6445 | -0.6612 |
| 0.0212 | 8.0 | 300 | 3.2506 | -8.9456 | -10.5410 | 0.5500 | 1.5954 | -243.8874 | -259.5571 | -0.2209 | -0.2378 |
| 0.0007 | 10.67 | 400 | 3.7538 | -0.8863 | -2.6713 | 0.4900 | 1.7850 | -235.1433 | -250.6024 | -0.7956 | -0.7980 |
| 0.0 | 13.33 | 500 | 3.7273 | -0.8791 | -2.4272 | 0.5100 | 1.5481 | -234.8721 | -250.5944 | -0.7748 | -0.7757 |
| 0.0 | 16.0 | 600 | 3.7631 | -0.9191 | -2.4137 | 0.4900 | 1.4946 | -234.8570 | -250.6388 | -0.7744 | -0.7750 |
| 0.0 | 18.67 | 700 | 3.7702 | -0.9575 | -2.4027 | 0.5100 | 1.4452 | -234.8448 | -250.6815 | -0.7745 | -0.7754 |
| 0.0 | 21.33 | 800 | 3.7953 | -0.9665 | -2.4000 | 0.5 | 1.4335 | -234.8418 | -250.6915 | -0.7742 | -0.7749 |
| 0.0 | 24.0 | 900 | 3.7744 | -0.9435 | -2.4093 | 0.5200 | 1.4658 | -234.8522 | -250.6659 | -0.7744 | -0.7751 |
| 0.0 | 26.67 | 1000 | 3.7428 | -0.9061 | -2.4348 | 0.5100 | 1.5287 | -234.8804 | -250.6243 | -0.7744 | -0.7754 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.1
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Library
- peft
- Created
- Apr 16, 2024
- Updated
- Apr 16, 2024