About fast_neural_style
TOUBH opened this issue · comments
How many rounds did you train in the fast neural style transfer experiment? I operate according to your steps, but the effect of the model I trained is not as good as the model you provided, and why is the model file I trained less than the file you provided by 3kb? I would like to know the reason and look forward to your reply!
And why is the suffix of the model I trained named model and the suffix of the model I provided named pth? As shown below, mosaic.pth was provided by the author, and painting.model was trained by me. I want to know why the sizes and suffixes of these two files are different because my training model does not work well
Was there an answer to this question?