doplhin-mistral-dpo-ultrafeedback-binarized-preferences-kto_pair
This model is a fine-tuned version of cognitivecomputations/dolphin-2.1-mistral-7b on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2024
- Rewards/chosen: 3.9237
- Rewards/rejected: -1.2199
- Rewards/accuracies: 0.7871
- Rewards/margins: 5.1437
- Logps/rejected: -319.8331
- Logps/chosen: -311.9210
- Logits/rejected: -2.2827
- Logits/chosen: -2.4135
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-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.2261 | 0.25 | 700 | 0.2691 | 3.8384 | 0.5201 | 0.6951 | 3.3183 | -302.4326 | -312.7741 | -2.4004 | -2.5103 |
| 0.1934 | 0.51 | 1400 | 0.2197 | 3.0918 | -2.0543 | 0.7802 | 5.1461 | -328.1768 | -320.2403 | -2.3140 | -2.4316 |
| 0.2163 | 0.76 | 2100 | 0.2024 | 3.9237 | -1.2199 | 0.7871 | 5.1437 | -319.8331 | -311.9210 | -2.2827 | -2.4135 |
Framework versions
- PEFT 0.8.2
- Transformers 4.37.2
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.2
- Downloads last month
- 2
Model tree for DrishtiSharma/dolphin-mistral-dpo-ultrafeedback-binarized-preferences-kto_pair
Base model
cognitivecomputations/dolphin-2.1-mistral-7b