metadata
base_model: microsoft/Phi-3-mini-4k-instruct
library_name: peft
license: mit
tags:
- trl
- sft
- generated_from_trainer
model-index:
- name: phi-3-mini-QLoRA
results: []
pipeline_tag: text-generation
phi-3-mini-QLoRA
This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1546
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.2359 | 0.1416 | 50 | 1.9291 |
| 1.5215 | 0.2833 | 100 | 1.2761 |
| 1.287 | 0.4249 | 150 | 1.2268 |
| 1.2613 | 0.5666 | 200 | 1.2073 |
| 1.2341 | 0.7082 | 250 | 1.1996 |
| 1.2037 | 0.8499 | 300 | 1.1953 |
| 1.2117 | 0.9915 | 350 | 1.1871 |
| 1.2023 | 1.1331 | 400 | 1.1813 |
| 1.1635 | 1.2748 | 450 | 1.1770 |
| 1.1689 | 1.4164 | 500 | 1.1732 |
| 1.2013 | 1.5581 | 550 | 1.1720 |
| 1.1853 | 1.6997 | 600 | 1.1675 |
| 1.1933 | 1.8414 | 650 | 1.1648 |
| 1.1774 | 1.9830 | 700 | 1.1619 |
| 1.1633 | 2.1246 | 750 | 1.1626 |
| 1.1756 | 2.2663 | 800 | 1.1593 |
| 1.1597 | 2.4079 | 850 | 1.1587 |
| 1.1599 | 2.5496 | 900 | 1.1562 |
| 1.1145 | 2.6912 | 950 | 1.1567 |
| 1.162 | 2.8329 | 1000 | 1.1553 |
| 1.1507 | 2.9745 | 1050 | 1.1546 |
Framework versions
- PEFT 0.12.0
- Transformers 4.43.3
- Pytorch 2.2.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1