metadata
base_model: mistralai/Mistral-7B-Instruct-v0.3
library_name: peft
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: >-
pgd_mistral_8bits_lr3.0000000000000004e-05_bs2_ep5_alpha32_rk4_do0.3_wd3.0e-04
results: []
pgd_mistral_8bits_lr3.0000000000000004e-05_bs2_ep5_alpha32_rk4_do0.3_wd3.0e-04
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9577
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: 3.0000000000000004e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 20
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.8635 | 0.9778 | 33 | 2.2032 |
| 1.6403 | 1.9852 | 67 | 1.2905 |
| 1.0775 | 2.9926 | 101 | 1.0257 |
| 0.9356 | 4.0 | 135 | 0.9717 |
| 0.9174 | 4.8889 | 165 | 0.9577 |
Framework versions
- PEFT 0.12.0
- Transformers 4.42.4
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1