metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: TheBloke/Mistral-7B-Instruct-v0.2-GPTQ
model-index:
- name: flippa-v1
results: []
flippa-v1
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: 0.8477
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: 0.0002
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- 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 |
|---|---|---|---|
| 1.2214 | 1.0 | 187 | 1.1960 |
| 1.0247 | 2.0 | 375 | 1.1154 |
| 0.9636 | 3.0 | 562 | 1.0554 |
| 0.9073 | 4.0 | 750 | 1.0037 |
| 0.8654 | 5.0 | 937 | 0.9583 |
| 0.8209 | 6.0 | 1125 | 0.9223 |
| 0.7994 | 7.0 | 1312 | 0.8916 |
| 0.7553 | 8.0 | 1500 | 0.8685 |
| 0.7415 | 9.0 | 1687 | 0.8537 |
| 0.724 | 9.97 | 1870 | 0.8477 |
Framework versions
- PEFT 0.9.0
- Transformers 4.38.1
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2