search

HasnainRaz / FC-DenseNet-TensorFlow

Fully Convolutional DenseNet (A.K.A 100 layer tiramisu) for semantic segmentation of images implemented in TensorFlow.
MIT License
123 stars 41 forks source link

about testing #8

Closed Leo551 closed 6 years ago

Leo551 commented 6 years ago

Thanks for your work. I have a question that after training how to test the model on my own data sets?

HasnainRaz commented 6 years ago

You can set the mode flag to infer and provide the test images directory path like so:

python main.py --mode=infer --infer_data=path/to/infer/images --output_folder=path/to/where/you/want/the/predicted/images/to/be/saved

You might have to modify the infer function in model.py, since currently it is a barebones predict function.