|
--- |
|
base_model: microsoft/Phi-3.5-mini-instruct |
|
library_name: peft |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: phi3.5-mini-adapter_v1 |
|
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. --> |
|
|
|
# phi3.5-mini-adapter_v1 |
|
|
|
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: 0.0998 |
|
|
|
## 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: 3e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_ratio: 0.05 |
|
- num_epochs: 50 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 14.2062 | 0.6061 | 10 | 13.2015 | |
|
| 5.0286 | 1.2121 | 20 | 3.9207 | |
|
| 0.248 | 1.8182 | 30 | 0.2396 | |
|
| 0.1801 | 2.4242 | 40 | 0.1860 | |
|
| 0.1496 | 3.0303 | 50 | 0.1639 | |
|
| 0.212 | 3.6364 | 60 | 0.1333 | |
|
| 0.0822 | 4.2424 | 70 | 0.1134 | |
|
| 0.07 | 4.8485 | 80 | 0.1061 | |
|
| 0.0871 | 5.4545 | 90 | 0.1178 | |
|
| 0.0645 | 6.0606 | 100 | 0.1017 | |
|
| 0.0558 | 6.6667 | 110 | 0.0998 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.11.1 |
|
- Transformers 4.43.1 |
|
- Pytorch 2.4.0+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
