Metadata-Version: 2.1
Name: tgim
Version: 0.0.0
Summary: tinygrad Image Models
Author: wozeparrot
License: Apache-2.0
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: tinygrad

# tinygrad Image Models

A collection of vision models implemented in tinygrad, in a similar vein to [timm](https://github.com/huggingface/pytorch-image-models).

Mostly targeting models trained on Imagenet-1k, and other models that are fast on resource-constrained devices.

## Models

- ShuffleNetV2 - [paper](https://arxiv.org/abs/1807.11164) [code](/tgim/models/shufflenetv2.py)
- GhostNetV2 - [paper](https://arxiv.org/abs/2211.12905) [code](/tgim/models/ghostnetv2.py)
- FocalNet - [paper](https://arxiv.org/abs/2203.11926) [code](/tgim/models/focalnet.py)
- FastViT - [paper](https://arxiv.org/abs/2303.14189) [code](/tgim/models/fastvit.py)
- RepViT - [paper](https://arxiv.org/abs/2307.09283) [code](/tgim/models/repvit.py)

## TODO

- [ ] For models that can be reparameterized, add that functionality
- [ ] Training

## License

See [LICENSE](/LICENSE).

Certain parts of the code are adapted from the original implementations, but they should all be under permissive licenses.
