How to make predictions using a trained model on my own dataset after training the model?

[halomei]

By running this command, you can input images from a specified directory and output the probabilities of each image under their respective labels, and then output the label with the highest probability.

[hankyul2]

Hi @halomei

You can infer images from a specific directory using a trained model by inference.py. The below script gives you example usage:

python3 inference.py --data-dir <your_image_folder> --split <split_folder_name> --model <your_model_name> --checkpoint <your_checkpoint_path> --mean <your_dataset_mean> --std <your_dataset_std>

I hope this will help you.

Thank you.

Hankyul.