A package to easily train powerful image classification models using colab's free TPUs.
Project description
The AI team at Décathlon Canada developed a library to help with the training of image classification models. It is specially made to exploit the free TPUs that are offered in Google colab notebooks. You can find the full documentation here
## Version 1.4.0
Added multilabel classification
## Version 1.3.0
Use suggested image sizes
Add EfficientNetV2 models
## Version 1.2.1
Change to Tensorflow 2.5
## Version 1.2.0
Add semi-supervised learning features
## Version 1.1.3
Link to public repository
## Version 1.1.2
Change name of package
## Version 1.1.1
Fix typo in split_train
## Version 1.1.0
Remove google scraping
## Version 1.0.0
Original release
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
decavision-1.4.0.tar.gz
(30.8 kB
view hashes)
Built Distribution
decavision-1.4.0-py3-none-any.whl
(33.0 kB
view hashes)
Close
Hashes for decavision-1.4.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | a6807f85654e50da59060557052197908e364c2c52789799558ed273721ac427 |
|
MD5 | 0064378c82fc0e5b5840829596011b62 |
|
BLAKE2b-256 | 01a161208b68a20f808137dbd253e0180b15b937df478af6fff2544017f960ec |