search

tyiannak / multimodal_movie_analysis

A Python Library for Multimodal Analysis of Movies and Content-based Movie Recommendation
25 stars 8 forks source link

Wrapper for classification #38 #39

Closed apoman38 closed 3 years ago

apoman38 commented 3 years ago

What this PR does:

Added a new file wrapper.py. This file receives an mp4 file as input and returns in which class it belongs to.

In the train.py file removed the one-hot encoding of the labeled data and added the saving of the trained model.

In the _analyzevisual.py added saving feature extraction in .npy format (not only in .csv).

How to QA:

Run the script python3 wrapper.py -f dataset/trump.mp4 -m SVM where -f option is the name of the file and -m the saved model of the training algorithm. The script must return the class that the file belongs.

tyiannak commented 3 years ago

So I guess I need to retrain the model as well right?

apoman38 commented 3 years ago

Yes right, because in the previous version train.py didn't save models.