gemma-2-9b-it-simpo-split-10-train_filtered_full
This model is a fine-tuned version of google/gemma-2-9b-it on the jz666/gemma2-ultrafeedback-ppl-split-10 dataset. It achieves the following results on the evaluation set:
- Loss: 2.7053
- Rewards/chosen: -19.9733
- Rewards/rejected: -25.3604
- Rewards/accuracies: 0.7746
- Rewards/margins: 5.3871
- Logps/rejected: -2.5360
- Logps/chosen: -1.9973
- Logits/rejected: -14.2055
- Logits/chosen: -14.1364
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: 8e-07
- train_batch_size: 2
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 16
- total_train_batch_size: 128
- total_eval_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: 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2.7717 | 0.8724 | 400 | 2.7137 | -19.8216 | -25.1579 | 0.7746 | 5.3363 | -2.5158 | -1.9822 | -14.1319 | -14.0744 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.7.0+cu128
- Datasets 2.18.0
- Tokenizers 0.19.1
- Downloads last month
- 31