metadata
license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
- trl
- sft
- generated_from_trainer
datasets:
- vipinkatara/SFT_data223
model-index:
- name: zephyr-7b-sft-full-10ep
results: []
zephyr-7b-sft-full-10ep
This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the vipinkatara/SFT_data223 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0485
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 64
- 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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.203 | 1.0 | 23 | 0.1086 |
| 0.0942 | 2.0 | 46 | 0.0702 |
| 0.064 | 3.0 | 69 | 0.0579 |
| 0.055 | 4.0 | 92 | 0.0539 |
| 0.0533 | 5.0 | 115 | 0.0529 |
| 0.0525 | 6.0 | 138 | 0.0515 |
| 0.0519 | 7.0 | 161 | 0.0505 |
| 0.0505 | 8.0 | 184 | 0.0496 |
| 0.0494 | 9.0 | 207 | 0.0487 |
| 0.0484 | 10.0 | 230 | 0.0485 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1