Skip to content

omermazig/DL-Workshop

Repository files navigation

NBA Shot Classifier Demo

Getting a dataset

You can run the create_videos_bank.ipynb notebook to create it yourself, but there might be some dependency management required for that. For example, py-tesseract needs you to first manually install the program manually

If you don't want to deal with that, you can download a tiny dataset (111 MB) from my drive, rename it dataset, and put it in the project's root folder. This of course will not be enough to train the model, but you'll be able to run inference.

Checking dataset metrics

Once you have a dataset, just run the check_model_inference.ipynb notebook, and you will get the full classification report. This could take a while, so you might want to start with a small dataset.

Predicting a single video

You can take any single video from the dataset, and upload it to the dedicated space on huggingface. You should get a prediction within about 30 seconds. There's also a space for multi label prediction, but as stated in Project.pdf, it is not as accurate.

This spaces should also support downloaded videos from YouTube (Tested it on a few and got good results), so you can also try that.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors