Cannot run train - Githubissues

ErisGe commented 1 year ago

I used the command python train_uapi.py --direc 'ixi/' to run train_uapi.py, but one mistake happened.

File "/home/hg/anaconda3/envs/SHD/lib/python3.8/site-packages/tensorflow/python/autograph/impl/api.py", line 692, in wrapper raise e.ag_error_metadata.to_exception(e) ValueError: Exception encountered when calling layer "gradient_penalty" (type GradientPenalty).

in user code:

File "/home/hg/unsupervised-image-transfer-and-uq-master/functions.py", line 66, in call  *
    norm_grad = K.sqrt(K.sum(K.square(gradients), axis=[1,2,3]))
File "/home/hg/anaconda3/envs/SHD/lib/python3.8/site-packages/keras/backend.py", line 2750, in square
    return tf.square(x)

ValueError: Tried to convert 'x' to a tensor and failed. Error: None values not supported.

Call arguments received: • inputs=['tf.Tensor(shape=(None, 5, 5, 1), dtype=float32)', 'tf.Tensor(shape=(None, 128, 128, 1), dtype=float32)']

anger-man commented 1 year ago

Hi ErisGe Unfortunately I was not able to reproduce this error. It seems that there have been a problem with the images in your case. Following things could rise this error:

corrupt .jpg data
an empty 'ixi/' directory or the directory is not following the structure described in the instructions
failed import of .jpg images. In the implementation, the command 'matplotlib.pyplot.imread' is used. Maybe check this.

Best, Christoph

ErisGe commented 1 year ago

Thanks! I will check these points.

anger-man / unsupervised-image-transfer-and-uq

Cannot run train #1