Type mismatch in fc_mask2 and self.out multiplication #15
Description
After the update last week, I ran into this issue. After reading the issues, the author fixed the typo on self.fc_mask2. However, the type is mismatched now.
self.fc_mask2 = tf.compat.v1.placeholder(tf.float32, shape=[None, self.num_Category], name='mask2') shape (?, 230)
self.out = tf.reshape(out, [-1, self.num_Event, self.num_Category]) shape (?, 2, 230)
Because of the shape mismatch, tensorflow can not do the matrix multiplication. It runs fine with the METABRIC dataset, which is composed of single risks. However, when running the code on the SYNTHETIC dataset and running the summarize_results.py, the program breaks.
Any solutions to fix this?