metadata
tags:
- trl
- dpo
- generated_from_trainer
model-index:
- name: zephyr-7b-dpo-full
results: []
zephyr-7b-dpo-full
This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0515
- Rewards/chosen: -0.5698
- Rewards/rejected: -1.0751
- Rewards/accuracies: 0.7305
- Rewards/margins: 0.5053
- Logps/rejected: -473.2190
- Logps/chosen: -416.4288
- Logits/rejected: 0.8846
- Logits/chosen: 0.7434
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: 4
- eval_batch_size: 8
- seed: 1
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- total_eval_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 |
---|---|---|---|---|---|---|---|---|---|---|---|
0.0993 | 0.21 | 100 | 0.1018 | -0.0745 | -0.2470 | 0.6875 | 0.1725 | -390.4119 | -366.9025 | 0.5707 | 0.5130 |
0.0553 | 0.42 | 200 | 0.0540 | -0.5063 | -0.9361 | 0.7305 | 0.4298 | -459.3217 | -410.0835 | 0.8994 | 0.7755 |
0.0503 | 0.63 | 300 | 0.0531 | -0.5353 | -1.0266 | 0.7461 | 0.4913 | -468.3688 | -412.9813 | 0.8095 | 0.6691 |
0.0537 | 0.84 | 400 | 0.0515 | -0.5698 | -1.0751 | 0.7305 | 0.5053 | -473.2190 | -416.4288 | 0.8846 | 0.7434 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1