Downloads · 30 days
3
10% of all-time downloads
model_hh_usp2_dpo1
guoyu-zhang/model_hh_usp2_dpo1
model_hh_usp2_dpo1 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 17, 2024
- Updated
- Apr 17, 2024
- SHA
- 1d7c0c6f
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_usp2_dpo1
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: 1.5291
- Rewards/chosen: -11.2131
- Rewards/rejected: -13.0930
- Rewards/accuracies: 0.6500
- Rewards/margins: 1.8799
- Logps/rejected: -246.0494
- Logps/chosen: -224.9259
- Logits/rejected: -0.9959
- Logits/chosen: -1.0145
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0903 | 2.67 | 100 | 0.9847 | -6.9568 | -7.7597 | 0.6300 | 0.8029 | -192.7170 | -182.3627 | -1.2298 | -1.2230 |
| 0.0022 | 5.33 | 200 | 1.3545 | -8.2875 | -9.7332 | 0.6000 | 1.4457 | -212.4518 | -195.6697 | -1.1216 | -1.0996 |
| 0.0001 | 8.0 | 300 | 1.4929 | -10.8994 | -12.7252 | 0.6500 | 1.8258 | -242.3715 | -221.7884 | -1.0149 | -1.0300 |
| 0.0001 | 10.67 | 400 | 1.5066 | -11.0254 | -12.8797 | 0.6500 | 1.8543 | -243.9168 | -223.0491 | -1.0069 | -1.0229 |
| 0.0001 | 13.33 | 500 | 1.5250 | -11.1335 | -12.9833 | 0.6500 | 1.8497 | -244.9522 | -224.1301 | -1.0009 | -1.0186 |
| 0.0001 | 16.0 | 600 | 1.5230 | -11.1873 | -13.0593 | 0.6500 | 1.8720 | -245.7126 | -224.6674 | -0.9979 | -1.0161 |
| 0.0001 | 18.67 | 700 | 1.5290 | -11.2177 | -13.0897 | 0.6500 | 1.8719 | -246.0163 | -224.9722 | -0.9960 | -1.0141 |
| 0.0001 | 21.33 | 800 | 1.5255 | -11.2152 | -13.0927 | 0.6500 | 1.8775 | -246.0468 | -224.9469 | -0.9960 | -1.0145 |
| 0.0001 | 24.0 | 900 | 1.5327 | -11.2235 | -13.0940 | 0.6500 | 1.8706 | -246.0598 | -225.0293 | -0.9961 | -1.0143 |
| 0.0001 | 26.67 | 1000 | 1.5291 | -11.2131 | -13.0930 | 0.6500 | 1.8799 | -246.0494 | -224.9259 | -0.9959 | -1.0145 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Library
- peft
- Created
- Apr 17, 2024
- Updated
- Apr 17, 2024