metadata
license: mit
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: microsoft/Phi-3-mini-128k-instruct
model-index:
- name: phi-3-mini-LoRA
results: []
phi-3-mini-LoRA
This model is a fine-tuned version of microsoft/Phi-3-mini-128k-instruct on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2692
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 |
|---|---|---|---|
| 1.7153 | 0.1766 | 100 | 1.4475 |
| 1.429 | 0.3532 | 200 | 1.4079 |
| 1.3977 | 0.5298 | 300 | 1.3931 |
| 1.3835 | 0.7064 | 400 | 1.3778 |
| 1.3612 | 0.8830 | 500 | 1.3640 |
| 1.3483 | 1.0596 | 600 | 1.3483 |
| 1.3115 | 1.2362 | 700 | 1.3340 |
| 1.2945 | 1.4128 | 800 | 1.3180 |
| 1.2829 | 1.5894 | 900 | 1.3074 |
| 1.2691 | 1.7660 | 1000 | 1.2954 |
| 1.26 | 1.9426 | 1100 | 1.2865 |
| 1.2259 | 2.1192 | 1200 | 1.2805 |
| 1.2117 | 2.2958 | 1300 | 1.2761 |
| 1.2001 | 2.4724 | 1400 | 1.2722 |
| 1.2035 | 2.6490 | 1500 | 1.2697 |
| 1.1836 | 2.8256 | 1600 | 1.2692 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1