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
For the model size and flops, you can calculate them precisely based on the sum of all basic operations.
For the memory, you can record the memory occupation of GPUs (use nvidia-smi to check).
Also, some packages can make this process easier, such as thop.
Hi,
May I kindly know how to calculate the model sizes/ memory/flops reported in Table 2 and 3 ?
Can you please share the code snippet ?
Thanks!
The text was updated successfully, but these errors were encountered: