---
license: mit
base_model: microsoft/Phi-3-mini-4k-instruct
tags:
- generated_from_trainer
model-index:
- name: Phi0503MA2
  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. -->

# Phi0503MA2

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: 0.0832

## 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 |
|:-------------:|:-----:|:----:|:---------------:|
| 3.7578        | 0.09  | 10   | 0.9244          |
| 0.4383        | 0.18  | 20   | 0.1617          |
| 0.1537        | 0.27  | 30   | 0.1389          |
| 0.1365        | 0.36  | 40   | 0.1165          |
| 0.1071        | 0.45  | 50   | 0.0962          |
| 0.1021        | 0.54  | 60   | 0.0964          |
| 0.0866        | 0.63  | 70   | 0.0848          |
| 0.0997        | 0.73  | 80   | 0.0891          |
| 0.08          | 0.82  | 90   | 0.0861          |
| 0.0813        | 0.91  | 100  | 0.0706          |
| 0.0675        | 1.0   | 110  | 0.0656          |
| 0.0626        | 1.09  | 120  | 0.0832          |
| 0.0641        | 1.18  | 130  | 0.0733          |
| 0.0693        | 1.27  | 140  | 0.0679          |
| 0.055         | 1.36  | 150  | 0.0745          |
| 0.0572        | 1.45  | 160  | 0.0603          |
| 0.0487        | 1.54  | 170  | 0.0614          |
| 0.0501        | 1.63  | 180  | 0.0610          |
| 0.0456        | 1.72  | 190  | 0.0660          |
| 0.0496        | 1.81  | 200  | 0.0626          |
| 0.0415        | 1.9   | 210  | 0.0626          |
| 0.0463        | 1.99  | 220  | 0.0663          |
| 0.0241        | 2.08  | 230  | 0.0741          |
| 0.0216        | 2.18  | 240  | 0.0932          |
| 0.0186        | 2.27  | 250  | 0.0979          |
| 0.0166        | 2.36  | 260  | 0.0852          |
| 0.0207        | 2.45  | 270  | 0.0819          |
| 0.0148        | 2.54  | 280  | 0.0857          |
| 0.0189        | 2.63  | 290  | 0.0866          |
| 0.0226        | 2.72  | 300  | 0.0844          |
| 0.0229        | 2.81  | 310  | 0.0841          |
| 0.02          | 2.9   | 320  | 0.0833          |
| 0.0187        | 2.99  | 330  | 0.0832          |


### Framework versions

- Transformers 4.36.0.dev0
- Pytorch 2.1.2+cu121
- Datasets 2.14.6
- Tokenizers 0.14.1