Downloads · 30 days
7
21% of all-time downloads
model_shp1_dpo5
guoyu-zhang/model_shp1_dpo5
model_shp1_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
33
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
- d189211f
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_shp1_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: 2.3217
- Rewards/chosen: -0.2355
- Rewards/rejected: -0.1439
- Rewards/accuracies: 0.5
- Rewards/margins: -0.0916
- Logps/rejected: -232.6779
- Logps/chosen: -231.7293
- Logits/rejected: -0.7538
- Logits/chosen: -0.6906
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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.0133 | 2.67 | 100 | 1.4101 | 1.9038 | 1.9347 | 0.4900 | -0.0309 | -228.5207 | -227.4508 | -0.7876 | -0.7426 |
| 0.0129 | 5.33 | 200 | 1.8574 | -0.4863 | -1.0084 | 0.5200 | 0.5221 | -234.4071 | -232.2310 | -0.7312 | -0.6995 |
| 0.0238 | 8.0 | 300 | 2.0008 | -1.7671 | -1.9006 | 0.5500 | 0.1335 | -236.1914 | -234.7925 | -0.5278 | -0.4564 |
| 0.0 | 10.67 | 400 | 2.3318 | -0.1632 | -0.0429 | 0.5 | -0.1202 | -232.4761 | -231.5847 | -0.7531 | -0.6899 |
| 0.0 | 13.33 | 500 | 2.3339 | -0.2064 | -0.0831 | 0.4900 | -0.1232 | -232.5565 | -231.6711 | -0.7531 | -0.6898 |
| 0.0 | 16.0 | 600 | 2.3248 | -0.2171 | -0.1093 | 0.5 | -0.1078 | -232.6089 | -231.6927 | -0.7533 | -0.6905 |
| 0.0 | 18.67 | 700 | 2.3383 | -0.2611 | -0.1205 | 0.5 | -0.1407 | -232.6312 | -231.7806 | -0.7531 | -0.6900 |
| 0.0 | 21.33 | 800 | 2.3166 | -0.2398 | -0.1509 | 0.5 | -0.0890 | -232.6919 | -231.7380 | -0.7534 | -0.6906 |
| 0.0 | 24.0 | 900 | 2.3279 | -0.2442 | -0.1312 | 0.5 | -0.1130 | -232.6526 | -231.7468 | -0.7537 | -0.6908 |
| 0.0 | 26.67 | 1000 | 2.3217 | -0.2355 | -0.1439 | 0.5 | -0.0916 | -232.6779 | -231.7293 | -0.7538 | -0.6906 |
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