phi-2-gpo-test-longest-iter-4
This model is a fine-tuned version of DUAL-GPO/phi-2-gpo-test-longest-iter-3 on the HuggingFaceH4/ultrafeedback_binarized dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0107
- Rewards/chosen: -0.0000
- Rewards/rejected: -0.0005
- Rewards/accuracies: 0.5085
- Rewards/margins: 0.0005
- Logps/rejected: -278.6688
- Logps/chosen: -306.3621
- Logits/rejected: 0.0917
- Logits/chosen: -0.0055
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: 5e-06
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- 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_ratio: 0.1
- num_epochs: 2
Training results
| Training Loss |
Epoch |
Step |
Validation Loss |
Rewards/chosen |
Rewards/rejected |
Rewards/accuracies |
Rewards/margins |
Logps/rejected |
Logps/chosen |
Logits/rejected |
Logits/chosen |
| 0.0105 |
1.6 |
100 |
0.0108 |
0.0006 |
0.0007 |
0.4945 |
-0.0001 |
-278.5431 |
-306.2985 |
0.0955 |
-0.0024 |
Framework versions
- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.2.1+cu121
- Datasets 2.14.6
- Tokenizers 0.15.2
