Open NoraHit opened 1 year ago
gtrain_output_tag = np.ones((batch_size_train, block_size[0] 3, 1)) label_smoothing_factor gtest_output_tag = np.ones((batch_size_test, block_size[0] 3, 1)) label_smoothing_factor
where does this block_size[0] come from??
Hi, we split the image volume into smaller blocks to comply with the GPU memory limit. In our study block_size = [64,64,64].
gtrain_output_tag = np.ones((batch_size_train, block_size[0] 3, 1)) label_smoothing_factor gtest_output_tag = np.ones((batch_size_test, block_size[0] 3, 1)) label_smoothing_factor