search

samaonline / Hierarchical-Model-for-Long-term-Video-Prediction

Given the previous frames of the video as input, we want to get the long-term frame prediction.
Apache License 2.0
31 stars 7 forks source link

analogy network problem #5

Open Katreenlooper opened 4 years ago

Katreenlooper commented 4 years ago

Hey, I am trying to run analogy network but squat folder is not exist (just squat_labels and squat_labels multiple).

Create dictionary of squat videos. Key = video number, Value = list of frames (numpy array images) of video videos = {} for video in os.listdir('squats/'): videos[video] = [] for frame in os.listdir('squats/' + str(video) + '/'): filename = 'squats/' + str(video) + '/' + str(frame) videos[video].append(imread(filename))

One thing, in your instructions you said that we should rename vgg16 to vgg, but in analogy network script (generator.init_weights) it is written vgg16.. so i modify it.

generator.init_weights(sess, 'models/alexnet.npy', 'models/vgg16.npy') discriminator.init_weights(sess, 'models/alexnet.npy', 'models/vgg16.npy')

Can you help me to overcome this ?

Thanks very much

samaonline commented 4 years ago

Hi, thanks much for your interest in the work. We have updated version here: https://github.com/ZhongxiaYan/video_prediction.

Can you try the new repo and let us know if there is additional question? thank you!

Katreenlooper commented 4 years ago

Hey, I checked that repo, and there is same problem. Where can i get squat directory in analogy_network.ipynb?

for video in os.listdir('squats/') videos[video] = [] for frame in os.listdir('squats/' + str(video) + '/'): filename = 'squats/' + str(video) + '/' + str(frame) videos[video].append(imread(filename))

Preprocess.ipynb generates two directories of (squat_labels, squat_labels_multiple) but there is no squat directory.. Thanks