metadata
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
- GaetanMichelet/chat-60_ft_task-1
library_name: peft
license: apache-2.0
tags:
- alignment-handbook
- trl
- sft
- generated_from_trainer
model-index:
- name: Mistral-7B_task-1_60-samples_config-1_full
results: []
Mistral-7B_task-1_60-samples_config-1_full
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the GaetanMichelet/chat-60_ft_task-1 dataset. It achieves the following results on the evaluation set:
- Loss: 0.9257
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: 0.0001
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 8
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.1428 | 0.8696 | 5 | 1.8872 |
| 1.7313 | 1.9130 | 11 | 1.4963 |
| 1.2743 | 2.9565 | 17 | 1.1005 |
| 0.9489 | 4.0 | 23 | 0.9686 |
| 0.7984 | 4.8696 | 28 | 0.9257 |
| 0.6358 | 5.9130 | 34 | 0.9723 |
| 0.4722 | 6.9565 | 40 | 1.0251 |
| 0.3258 | 8.0 | 46 | 1.1248 |
| 0.2366 | 8.8696 | 51 | 1.2418 |
| 0.1887 | 9.9130 | 57 | 1.3003 |
| 0.1497 | 10.9565 | 63 | 1.3586 |
| 0.1313 | 12.0 | 69 | 1.4447 |
Framework versions
- PEFT 0.12.0
- Transformers 4.44.0
- Pytorch 2.1.2+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1