pointer-generator
pointer-generator copied to clipboard
NAN source?
https://github.com/abisee/pointer-generator/blob/b29e986f24fdd01a6b6d6008187c5c887f0be282/attention_decoder.py#L101
return attn_dist / (tf.reshape(masked_sums, [-1, 1]) + tf.ones_like(tf.reshape(masked_sums, [-1, 1])) * sys.float_info.epsilon)
tf.reshape(masked_sums, [-1, 1]) seems will not be zeros during training, but it is also harmful due to "division by 0", NAN issue may be caused by this reason.