search

leeyeehoo / CSRNet-pytorch

CSRNet: Dilated Convolutional Neural Networks for Understanding the Highly Congested Scenes
647 stars 261 forks source link

Generatig ground truth .mat vs. .h5 files #69

Closed kherings closed 4 years ago

kherings commented 4 years ago

Your code tries to load .h5 files from the ground truth directory: with h5py.File(img_path.replace('.jpg','.h5').replace('images','ground_truth'), 'w')

Neither are .h5 files included in the google drive file, nor you do generate them in the code. (as far as I ca see it). Did you omit any intermediate steps that you performed elsewhere?

(Or do you actually load the .mat files to h5py?)

        image_object = bucket.Object(img_path.replace('.jpg','.mat').replace('images/','ground_truth/GT_'))    
        with h5py.File(image_object.get()['Body'].read(), 'w') as hf: :
            hf['density'] = k