Automatic-Image-Colorization
Automatic-Image-Colorization copied to clipboard
Armour
🎨 Automatic Image Colorization using TensorFlow based on Residual Encoder Network
Hello! I've found a performance issue in /read_input.py: `dataset.batch(batch_size=batch_size)`[(here)](https://github.com/Armour/Automatic-Image-Colorization/blob/fd59980929150ec5bd03cc6d9d23cd25eb35668a/read_input.py#L88) should be called before `dataset.map(read_image)`[(here)](https://github.com/Armour/Automatic-Image-Colorization/blob/fd59980929150ec5bd03cc6d9d23cd25eb35668a/read_input.py#L85), , which could make your program more efficient. Here is [the tensorflow document](https://tensorflow.google.cn/guide/data_performance?hl=zh_cn#vectorized_mapping) to support it....
Hello,I found a performance issue in the definition of `get_dataset_iterator` , read_input.py, [dataset = dataset.map(read_image)](https://github.com/Armour/Automatic-Image-Colorization/blob/fd59980929150ec5bd03cc6d9d23cd25eb35668a/read_input.py#L85) was called without **num_parallel_calls**. I think it will increase the efficiency of your program if...
while begin training ，the use of memory keeps growth until memory error。my memory is 8G
any suggsetion for train set format？only use color images？any preprocess of color images？thank you
…r to point to correct directory **Make sure the PR fulfills these requirements:** - When resolving a specific issue, make sure it's referenced in the PR's title (e.g. `Closes #xxx[,#xxx]`,...
It ain't much, but it's honest work **Make sure the PR fulfills these requirements:** - When resolving a specific issue, make sure it's referenced in the PR's title (e.g. `Closes...
