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: []
phi-3-mini-QLoRA
This model is a fine-tuned version of microsoft/Phi-3-mini-4k-instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5616
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- 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 |
|---|---|---|---|
| 0.8926 | 0.1809 | 100 | 0.6390 |
| 0.6079 | 0.3618 | 200 | 0.5857 |
| 0.5794 | 0.5427 | 300 | 0.5772 |
| 0.582 | 0.7237 | 400 | 0.5727 |
| 0.565 | 0.9046 | 500 | 0.5705 |
| 0.5773 | 1.0855 | 600 | 0.5685 |
| 0.5589 | 1.2664 | 700 | 0.5668 |
| 0.5611 | 1.4473 | 800 | 0.5657 |
| 0.5748 | 1.6282 | 900 | 0.5645 |
| 0.5594 | 1.8091 | 1000 | 0.5641 |
| 0.5607 | 1.9900 | 1100 | 0.5636 |
| 0.5504 | 2.1710 | 1200 | 0.5634 |
| 0.5663 | 2.3519 | 1300 | 0.5623 |
| 0.5542 | 2.5328 | 1400 | 0.5621 |
| 0.5549 | 2.7137 | 1500 | 0.5617 |
| 0.5577 | 2.8946 | 1600 | 0.5616 |
Framework versions
- PEFT 0.12.0
- Transformers 4.43.3
- Pytorch 2.3.1
- Datasets 2.20.0
- Tokenizers 0.19.1