https://github.com/mendhak/gpt2-custom-model-tutorial
Supporting setup and scripts for training your own custom model for GPT2
https://github.com/mendhak/gpt2-custom-model-tutorial
Last synced: 8 months ago
JSON representation
Supporting setup and scripts for training your own custom model for GPT2
- Host: GitHub
- URL: https://github.com/mendhak/gpt2-custom-model-tutorial
- Owner: mendhak
- Created: 2023-04-09T16:42:47.000Z (about 3 years ago)
- Default Branch: master
- Last Pushed: 2023-04-10T16:50:35.000Z (about 3 years ago)
- Last Synced: 2025-08-15T15:46:01.088Z (10 months ago)
- Language: Python
- Size: 21.5 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
Repository to support [a tutorial](https://towardsdatascience.com/train-gpt-2-in-your-own-language-fc6ad4d60171) on using GPT-2 to train your own model. Tested on Ubuntu 22.04 with NVidia RTX 2080 Ti.
## Setup
[Set up TensorFlow](https://www.tensorflow.org/install/pip#linux) by following the Step-by-step instructions, not the quick ones. Include steps to test the NVIDIA GPU and the fix for Ubuntu 22.04.
Repeated instructions:
```
conda env create -f environment.conda.yml python=3.9
conda activate tf
pip install -r requirements.txt
```
## Disk Usage
A lot of files get downloaded and consume disk space.
Transformers cache: ~/.cache/huggingface/
Anaconda: ~/anaconda3
Miniconda: ~/miniconda3