Training of CLIP in JAX
Project description
CLIP-JAX
This repository is used to CLIP models from 🤗 transformers using JAX.
Installation
pip install -e .
Usage
- Use dataset/prepare_dataset.ipynb to prepare your dataset.
- Train the model with training/train_clip.py.
Supported downstream tasks
- Image classification with
FlaxCLIPVisionModelForImageClassification
TODO
- Add guides
- Add pre-trained models
- Add more downstream tasks
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
clip_jax-0.0.1.post1.tar.gz
(24.0 kB
view hashes)
Built Distribution
Close
Hashes for clip_jax-0.0.1.post1-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 617316cee62eee57ff68e01962521f41f0be794950949cc935d0478acc7d0132 |
|
MD5 | 3b6ca0da0304c4b8e0fef1f3836844d0 |
|
BLAKE2b-256 | c697a0e6e1b6d617e88ab572acc604394034be70628c201719b4e5067f5673b8 |