multi_lable_multiclass_classification

Usage:

1. Preprocessing and training on GPU
   training notebook.
   

2. Inference File
   -> Run the script as illustrated in the notebook above (in last cell).
   

$ python inference.py --csv_file path_to_test_csv --ckpts path_to_model_weights

colab example given that you have added the inference file,test.csv from repo to colab and downloaded weights from link below or gdown --id 1-D4QzRDhtlFqj4wzZUj4s9TpS2xA369t

$ python inference.py --csv_file attributes_test.csv --ckpts model.pth

-> This will generate Output.csv file in the same directory as of inference.py

Models

Under the outputs directory, model weights and loss graph are available. Drive link to model : model.pt