simpo-lora-2
This model is a fine-tuned version of hatakeyama-llm-team/with_halcination_little_codes_ck5200 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2390
- Rewards/chosen: -4.5674
- Rewards/rejected: -5.4724
- Rewards/accuracies: 0.6150
- Rewards/margins: 0.9050
- Logps/rejected: -2.1889
- Logps/chosen: -1.8270
- Logits/rejected: 0.5316
- Logits/chosen: 0.4876
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: 1e-06
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 64
- total_train_batch_size: 64
- 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1.2463 | 0.1600 | 100 | 1.2850 | -4.5732 | -5.3937 | 0.5984 | 0.8205 | -2.1575 | -1.8293 | 0.6303 | 0.5626 |
| 1.3045 | 0.3200 | 200 | 1.2645 | -4.5737 | -5.4306 | 0.6014 | 0.8569 | -2.1722 | -1.8295 | 0.5868 | 0.5287 |
| 1.301 | 0.4800 | 300 | 1.2457 | -4.5684 | -5.4599 | 0.6111 | 0.8915 | -2.1840 | -1.8274 | 0.5387 | 0.4911 |
| 1.4689 | 0.6400 | 400 | 1.2386 | -4.5673 | -5.4725 | 0.6131 | 0.9052 | -2.1890 | -1.8269 | 0.5307 | 0.4859 |
| 1.3989 | 0.8000 | 500 | 1.2386 | -4.5680 | -5.4717 | 0.6121 | 0.9036 | -2.1887 | -1.8272 | 0.5340 | 0.4899 |
| 1.1781 | 0.9600 | 600 | 1.2390 | -4.5674 | -5.4724 | 0.6150 | 0.9050 | -2.1889 | -1.8270 | 0.5316 | 0.4876 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
- Downloads last month
- 0