Is it possible to set training batchsize larger than 1?

[hot-dog]

No description provided.

[justinjohn0306]

Were you able to generate the results?

…

On Tue, Jan 19, 2021, 5:43 PM hot-dog ***@***.***> wrote: — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#13>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIDVKQ5CWDHRW36LCFCRX33S2VZP3ANCNFSM4WIV63ZA> .

[hot-dog]

I am trying with my own data, but the training speed is too slow, it still need some time.

[justinjohn0306]

ok well do let me know if you find any success

…

On Tue, Jan 19, 2021 at 6:02 PM hot-dog ***@***.***> wrote: I am trying with my own data, but the training speed is too slow, it still need some time. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#13 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIDVKQZ5CMAPLPSQ6I2BQGDS2V3VLANCNFSM4WIV63ZA> .

[carmelofascella]

I have tried to train using varying values of batchSize (2,3,4,8,12..) but I have always got this error during the first epoch:

Do you have any ideas about how to solve it?

[zhangyahu1]

Same here