metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
datasets:
- generator
base_model: mistralai/Mistral-7B-Instruct-v0.2
model-index:
- name: mistral_instruct_generation
results: []
mistral_instruct_generation
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 0.4281
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: 0.0003
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.8896 | 1.0 | 17 | 0.6754 |
| 0.5429 | 2.0 | 34 | 0.4964 |
| 0.4271 | 3.0 | 51 | 0.4475 |
| 0.386 | 4.0 | 68 | 0.4246 |
| 0.3377 | 5.0 | 85 | 0.4118 |
| 0.3068 | 6.0 | 102 | 0.4102 |
| 0.29 | 7.0 | 119 | 0.4088 |
| 0.2638 | 8.0 | 136 | 0.4112 |
| 0.2424 | 9.0 | 153 | 0.4268 |
| 0.2317 | 10.0 | 170 | 0.4281 |
Framework versions
- PEFT 0.10.0
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2