|
--- |
|
base_model: microsoft/Phi-3.5-mini-instruct |
|
library_name: peft |
|
license: mit |
|
tags: |
|
- trl |
|
- sft |
|
- generated_from_trainer |
|
model-index: |
|
- name: phi-3-mini-LoRA |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# phi-3-mini-LoRA |
|
|
|
This model is a fine-tuned version of [microsoft/Phi-3.5-mini-instruct](https://huggingface.co/microsoft/Phi-3.5-mini-instruct) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4630 |
|
|
|
## 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: 1e-05 |
|
- 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: 0.2 |
|
- training_steps: 250 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-------:|:----:|:---------------:| |
|
| 1.6578 | 1.3333 | 5 | 1.6863 | |
|
| 1.6277 | 2.6667 | 10 | 1.6771 | |
|
| 1.615 | 4.0 | 15 | 1.6615 | |
|
| 1.5879 | 5.3333 | 20 | 1.6372 | |
|
| 1.5835 | 6.6667 | 25 | 1.6028 | |
|
| 1.5908 | 8.0 | 30 | 1.5586 | |
|
| 1.5143 | 9.3333 | 35 | 1.5012 | |
|
| 1.4633 | 10.6667 | 40 | 1.4352 | |
|
| 1.3414 | 12.0 | 45 | 1.3606 | |
|
| 1.3229 | 13.3333 | 50 | 1.2811 | |
|
| 1.2218 | 14.6667 | 55 | 1.2119 | |
|
| 1.1352 | 16.0 | 60 | 1.1488 | |
|
| 1.0852 | 17.3333 | 65 | 1.0885 | |
|
| 0.9989 | 18.6667 | 70 | 1.0299 | |
|
| 0.9959 | 20.0 | 75 | 0.9757 | |
|
| 0.921 | 21.3333 | 80 | 0.9205 | |
|
| 0.8727 | 22.6667 | 85 | 0.8683 | |
|
| 0.8067 | 24.0 | 90 | 0.8200 | |
|
| 0.7785 | 25.3333 | 95 | 0.7783 | |
|
| 0.7139 | 26.6667 | 100 | 0.7396 | |
|
| 0.7081 | 28.0 | 105 | 0.7095 | |
|
| 0.6705 | 29.3333 | 110 | 0.6824 | |
|
| 0.6177 | 30.6667 | 115 | 0.6613 | |
|
| 0.6106 | 32.0 | 120 | 0.6418 | |
|
| 0.575 | 33.3333 | 125 | 0.6239 | |
|
| 0.5904 | 34.6667 | 130 | 0.6083 | |
|
| 0.5917 | 36.0 | 135 | 0.5927 | |
|
| 0.5051 | 37.3333 | 140 | 0.5801 | |
|
| 0.5169 | 38.6667 | 145 | 0.5656 | |
|
| 0.5442 | 40.0 | 150 | 0.5542 | |
|
| 0.5112 | 41.3333 | 155 | 0.5432 | |
|
| 0.5061 | 42.6667 | 160 | 0.5321 | |
|
| 0.5071 | 44.0 | 165 | 0.5234 | |
|
| 0.4373 | 45.3333 | 170 | 0.5119 | |
|
| 0.4476 | 46.6667 | 175 | 0.5049 | |
|
| 0.3914 | 48.0 | 180 | 0.4972 | |
|
| 0.465 | 49.3333 | 185 | 0.4914 | |
|
| 0.4122 | 50.6667 | 190 | 0.4890 | |
|
| 0.4209 | 52.0 | 195 | 0.4837 | |
|
| 0.3933 | 53.3333 | 200 | 0.4784 | |
|
| 0.3583 | 54.6667 | 205 | 0.4760 | |
|
| 0.3952 | 56.0 | 210 | 0.4727 | |
|
| 0.3858 | 57.3333 | 215 | 0.4708 | |
|
| 0.3433 | 58.6667 | 220 | 0.4707 | |
|
| 0.4041 | 60.0 | 225 | 0.4680 | |
|
| 0.3558 | 61.3333 | 230 | 0.4665 | |
|
| 0.382 | 62.6667 | 235 | 0.4650 | |
|
| 0.3625 | 64.0 | 240 | 0.4638 | |
|
| 0.3513 | 65.3333 | 245 | 0.4644 | |
|
| 0.3541 | 66.6667 | 250 | 0.4630 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.13.2 |
|
- Transformers 4.45.2 |
|
- Pytorch 2.4.0 |
|
- Datasets 3.0.1 |
|
- Tokenizers 0.20.0 |
