metadata
license: gemma
base_model: ale-bay/zephyr-2b-gemma-sft
tags:
- alignment-handbook
- trl
- dpo
- generated_from_trainer
- trl
- dpo
- generated_from_trainer
datasets:
- argilla/dpo-mix-7k
model-index:
- name: zephyr-2b-gemma-sft-dpo
results: []
zephyr-2b-gemma-sft-dpo
This model is a fine-tuned version of ale-bay/zephyr-2b-gemma-sft on the argilla/dpo-mix-7k dataset. It achieves the following results on the evaluation set:
- Loss: 0.5950
- Rewards/chosen: 0.2219
- Rewards/rejected: -0.0437
- Rewards/accuracies: 0.7188
- Rewards/margins: 0.2656
- Logps/rejected: -363.9149
- Logps/chosen: -363.8262
- Logits/rejected: -15.9487
- Logits/chosen: -16.5867
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-07
- train_batch_size: 2
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 8
- total_train_batch_size: 128
- total_eval_batch_size: 32
- 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.5725 | 1.8957 | 100 | 0.5960 | 0.2261 | -0.0336 | 0.7188 | 0.2597 | -363.7130 | -363.7419 | -15.9487 | -16.5860 |
Framework versions
- Transformers 4.42.3
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1