Closed bysowhat closed 4 years ago
hi, could you figure out what is the 'pool_every_timestep' value in original paper? thanks:)
Hello, I have the same doubt with you, In paper, author said "we provide the pooled context only once as input to the decoder", it is the opposite of "pool_every_timestep". You closed the issue so is it solved?
hi, could you figure out what is the 'pool_every_timestep' value in original paper? thanks:)