Downloads · 30 days
1
4% of all-time downloads
model_hh_shp3_dpo5
guoyu-zhang/model_hh_shp3_dpo5
model_hh_shp3_dpo5 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
28
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
- 8eac6302
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_shp3_dpo5
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: 3.2982
- Rewards/chosen: -7.3754
- Rewards/rejected: -9.0651
- Rewards/accuracies: 0.5900
- Rewards/margins: 1.6897
- Logps/rejected: -254.8305
- Logps/chosen: -235.6870
- Logits/rejected: -1.0713
- Logits/chosen: -1.0440
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0495 | 2.67 | 100 | 1.4386 | 0.3553 | -0.0876 | 0.5400 | 0.4429 | -236.8756 | -220.2257 | -0.5196 | -0.4726 |
| 0.0246 | 5.33 | 200 | 2.9978 | -3.8865 | -4.5338 | 0.5400 | 0.6473 | -245.7680 | -228.7092 | -0.6969 | -0.6601 |
| 0.0049 | 8.0 | 300 | 3.2131 | -5.2339 | -6.8669 | 0.5700 | 1.6330 | -250.4342 | -231.4039 | -0.9891 | -0.9685 |
| 0.0 | 10.67 | 400 | 3.2941 | -7.3754 | -9.0927 | 0.5900 | 1.7173 | -254.8858 | -235.6871 | -1.0720 | -1.0447 |
| 0.0 | 13.33 | 500 | 3.3095 | -7.4032 | -9.1002 | 0.5800 | 1.6970 | -254.9008 | -235.7425 | -1.0717 | -1.0441 |
| 0.0 | 16.0 | 600 | 3.2770 | -7.3721 | -9.0946 | 0.5900 | 1.7225 | -254.8896 | -235.6804 | -1.0714 | -1.0442 |
| 0.0 | 18.67 | 700 | 3.2884 | -7.3870 | -9.1102 | 0.5900 | 1.7231 | -254.9207 | -235.7102 | -1.0715 | -1.0443 |
| 0.0 | 21.33 | 800 | 3.3005 | -7.3860 | -9.0763 | 0.5900 | 1.6902 | -254.8528 | -235.7083 | -1.0716 | -1.0442 |
| 0.0 | 24.0 | 900 | 3.2838 | -7.3850 | -9.0845 | 0.5900 | 1.6994 | -254.8693 | -235.7063 | -1.0716 | -1.0441 |
| 0.0 | 26.67 | 1000 | 3.2982 | -7.3754 | -9.0651 | 0.5900 | 1.6897 | -254.8305 | -235.6870 | -1.0713 | -1.0440 |
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