Downloads · 30 days
3
11% of all-time downloads
model_hh_usp1_dpo9
guoyu-zhang/model_hh_usp1_dpo9
model_hh_usp1_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
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
- 7133c60c
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_usp1_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: 2.5218
- Rewards/chosen: 4.1305
- Rewards/rejected: -4.5499
- Rewards/accuracies: 0.6900
- Rewards/margins: 8.6805
- Logps/rejected: -117.5696
- Logps/chosen: -105.7146
- Logits/rejected: -0.1040
- Logits/chosen: -0.0322
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0799 | 2.67 | 100 | 2.1948 | 0.0613 | -2.4589 | 0.6400 | 2.5202 | -115.2462 | -110.2360 | -0.3648 | -0.3265 |
| 0.0245 | 5.33 | 200 | 3.0197 | -6.4081 | -13.9036 | 0.6600 | 7.4954 | -127.9625 | -117.4242 | -0.6260 | -0.5678 |
| 0.0588 | 8.0 | 300 | 2.7834 | -6.7237 | -15.1196 | 0.7200 | 8.3958 | -129.3136 | -117.7749 | -0.3355 | -0.2470 |
| 0.0 | 10.67 | 400 | 2.5742 | 5.1319 | -3.5242 | 0.6600 | 8.6561 | -116.4299 | -104.6020 | -0.1518 | -0.0842 |
| 0.0 | 13.33 | 500 | 2.5287 | 4.1796 | -4.4812 | 0.7100 | 8.6608 | -117.4932 | -105.6601 | -0.1019 | -0.0304 |
| 0.0 | 16.0 | 600 | 2.5247 | 4.1189 | -4.5458 | 0.7000 | 8.6646 | -117.5649 | -105.7276 | -0.1019 | -0.0304 |
| 0.0 | 18.67 | 700 | 2.5257 | 4.1387 | -4.5517 | 0.7000 | 8.6904 | -117.5715 | -105.7055 | -0.1024 | -0.0311 |
| 0.0 | 21.33 | 800 | 2.5440 | 4.1061 | -4.5342 | 0.7000 | 8.6404 | -117.5521 | -105.7417 | -0.1035 | -0.0320 |
| 0.0 | 24.0 | 900 | 2.5242 | 4.1345 | -4.5168 | 0.7100 | 8.6512 | -117.5327 | -105.7102 | -0.1029 | -0.0313 |
| 0.0 | 26.67 | 1000 | 2.5218 | 4.1305 | -4.5499 | 0.6900 | 8.6805 | -117.5696 | -105.7146 | -0.1040 | -0.0322 |
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 16, 2024
- Updated
- Apr 16, 2024