metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ
model-index:
- name: mistral-algerian-dialect-translation
results: []
mistral-algerian-dialect-translation
This model is a fine-tuned version of TheBloke/Mistral-7B-Instruct-v0.2-GPTQ on the None dataset. It achieves the following results on the evaluation set:
- Loss: 4.3666
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: 25
- eval_batch_size: 25
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 100
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.1728 | 0.93 | 10 | 5.5818 |
| 5.3777 | 1.95 | 21 | 5.3171 |
| 5.106 | 2.98 | 32 | 5.0503 |
| 4.8644 | 4.0 | 43 | 4.8203 |
| 5.1233 | 4.93 | 53 | 4.6552 |
| 4.5071 | 5.95 | 64 | 4.5252 |
| 4.4027 | 6.98 | 75 | 4.4408 |
| 4.34 | 8.0 | 86 | 4.3903 |
| 4.7357 | 8.93 | 96 | 4.3688 |
| 3.5274 | 9.3 | 100 | 4.3666 |
Framework versions
- PEFT 0.10.0
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2