metadata
base_model: mistralai/Mistral-7B-v0.1
tags:
- alignment-handbook
- generated_from_trainer
datasets:
- HuggingFaceH4/ultrafeedback_binarized
model-index:
- name: mistral_7b_gsm8k_ep2_1e-5_dpo
results: []
mistral_7b_gsm8k_ep2_1e-5_dpo
This model is a fine-tuned version of /home/hyeonbin/self_train/Verifiers/models/mistral_7b_gsm8k_ep2_1e-5_rft_round1 on the GSM8K Train Set. It achieves the following results on the evaluation set (=GSM8K Train subset):
- Loss: 0.0005
- Rewards/chosen: -1.7120
- Rewards/rejected: -14.3548
- Rewards/accuracies: 1.0
- Rewards/margins: 12.6428
- Logps/rejected: -1466.6733
- Logps/chosen: -208.2280
- Logits/rejected: -3.2168
- Logits/chosen: -2.3996
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-07
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.014 | 1.0 | 7066 | 0.0005 | -1.7120 | -14.3548 | 1.0 | 12.6428 | -1466.6733 | -208.2280 | -3.2168 | -2.3996 |
Framework versions
- Transformers 4.41.0
- Pytorch 2.3.0+cu121
- Datasets 2.14.6
- Tokenizers 0.19.1