# Transformer-Trainer

**Repository Path**: Jackson0829/Transformer-Trainer

## Basic Information

- **Project Name**: Transformer-Trainer
- **Description**: This demo shows how to use the Trainer module of HuggingFaces's Transformers
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-11-04
- **Last Updated**: 2021-11-04

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README



# This demo shows how to use the Trainer module of HuggingFaces's Transformers

### dataset: IMDB

You can download it from [here][http://ai.stanford.edu/~amaas/data/sentiment/]

+ Train data: 25000

  + pos:12500
  + neg:12500

  --> use 80% for Training and 20% for Dev

+ Test data: 25000

  + pos:12500
  + neg:12500

### File Structure

+ /model 
  + Contains Model file which uses Bert [cls] and a Linear layer
+ /result
  + Save checkpoint and result record
+ dataset.py
  + Data Preprocess and make it as Dataset of torch.utils.data
+ parser1.py
  + Input hyper parameters
+ Train.py
  + Training File