Open HaohanWang opened 2 years ago
maybe the code is only intended to be used when batch_size = 1 ?
Will take a look tonight and tmr before our meeting.
Daniel
On Nov 18, 2021, at 9:27 PM, Haohan Wang @.***> wrote:
maybe the code is only intended to be used when batch_size = 1 ?
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.
@gn03249822 The current code seems to be a factor of batch size, especially about the code taking the average of the embeddings / losses of the batch like here . This does not seem to make sense to me.