|
--- |
|
license: mit |
|
base_model: microsoft/Phi-3-mini-4k-instruct |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: Phi0503B1 |
|
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. --> |
|
|
|
# Phi0503B1 |
|
|
|
This model is a fine-tuned version of [microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0800 |
|
|
|
## 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: 0.0003 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 16 |
|
- total_train_batch_size: 128 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine_with_restarts |
|
- lr_scheduler_warmup_steps: 100 |
|
- num_epochs: 3 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 3.5697 | 0.09 | 10 | 0.7185 | |
|
| 0.345 | 0.18 | 20 | 0.1655 | |
|
| 0.1552 | 0.27 | 30 | 0.1343 | |
|
| 0.1345 | 0.36 | 40 | 0.1175 | |
|
| 0.121 | 0.45 | 50 | 0.1152 | |
|
| 0.1088 | 0.54 | 60 | 0.0861 | |
|
| 0.0923 | 0.63 | 70 | 0.0942 | |
|
| 0.0773 | 0.73 | 80 | 0.0681 | |
|
| 0.0606 | 0.82 | 90 | 0.0686 | |
|
| 0.0647 | 0.91 | 100 | 0.0624 | |
|
| 0.062 | 1.0 | 110 | 0.0663 | |
|
| 0.0434 | 1.09 | 120 | 0.0687 | |
|
| 0.042 | 1.18 | 130 | 0.0675 | |
|
| 0.0503 | 1.27 | 140 | 0.0681 | |
|
| 0.0445 | 1.36 | 150 | 0.0654 | |
|
| 0.0511 | 1.45 | 160 | 0.0593 | |
|
| 0.0462 | 1.54 | 170 | 0.0687 | |
|
| 0.0498 | 1.63 | 180 | 0.0651 | |
|
| 0.0448 | 1.72 | 190 | 0.0640 | |
|
| 0.043 | 1.81 | 200 | 0.0636 | |
|
| 0.04 | 1.9 | 210 | 0.0617 | |
|
| 0.043 | 1.99 | 220 | 0.0613 | |
|
| 0.0226 | 2.08 | 230 | 0.0657 | |
|
| 0.0165 | 2.18 | 240 | 0.0788 | |
|
| 0.011 | 2.27 | 250 | 0.0943 | |
|
| 0.0097 | 2.36 | 260 | 0.0946 | |
|
| 0.0167 | 2.45 | 270 | 0.0864 | |
|
| 0.0105 | 2.54 | 280 | 0.0827 | |
|
| 0.0118 | 2.63 | 290 | 0.0819 | |
|
| 0.0156 | 2.72 | 300 | 0.0802 | |
|
| 0.0137 | 2.81 | 310 | 0.0800 | |
|
| 0.013 | 2.9 | 320 | 0.0800 | |
|
| 0.0098 | 2.99 | 330 | 0.0800 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.36.0.dev0 |
|
- Pytorch 2.1.2+cu121 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.14.1 |
|
|
