You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Dear Author, I am really appreciated by your great works. Thanks a lot.
Now I have a questions of implementation on Algorithm,
Q1.
As you mentioned in the paper, if c is lower than r, then the current minibatch modality are added to the buffer for true negative.
then may I comprehend that there is other buffers, which is similar ones like queues for each modality?
Q2.
For determining whether the c_k of x_k and c_k of y_k are lower than the hyper parameter r,
if I detect the c_k of x_k and c_k of y_k among the queue_x and queue_y are truly lower than r,
how can I implement it in pytorch code?
Q3.
If you do not mind, I wanna get your full code, so, can I get the part of them if you are reluctant to send the full code?
I really feel thirsty and feel really really fascinated by your work.
If you can, please send me email : [email protected]
I DO promise not to distribute any other guys if there is no permission to.
Again, Thank you for your attention.
The text was updated successfully, but these errors were encountered:
Dear Author, I am really appreciated by your great works. Thanks a lot.
Now I have a questions of implementation on Algorithm,
Q1.
As you mentioned in the paper, if c is lower than r, then the current minibatch modality are added to the buffer for true negative.
then may I comprehend that there is other buffers, which is similar ones like queues for each modality?
Q2.
For determining whether the c_k of x_k and c_k of y_k are lower than the hyper parameter r,
if I detect the c_k of x_k and c_k of y_k among the queue_x and queue_y are truly lower than r,
how can I implement it in pytorch code?
Q3.
If you do not mind, I wanna get your full code, so, can I get the part of them if you are reluctant to send the full code?
I really feel thirsty and feel really really fascinated by your work.
If you can, please send me email : [email protected]
I DO promise not to distribute any other guys if there is no permission to.
Again, Thank you for your attention.
The text was updated successfully, but these errors were encountered: