Closed NewBornRustacean closed 4 months ago
Great work! It looks correct to me on first pass. Is there any way to implement it without requiring a 7D tensor? Like maybe merging some of the dimensions before doing the last pool_last_dim? Ideally we keep the max tensor dims at 6 (in order to keep the shapetracker small because it's stored on the stack).
If not, I'll look into adding a 7th dimension
Thanks! I' ll look into it! (I'm quite a newbie just like my nick name, this is really helpful! )
Hello! Thanks for this awesome project! @jafioti
Recently, I'm working on Conv3d(#28 #29) and I have a question:
My draft is here:
I'm not sure I'm going right direction, so any comments will be appreciated!
Have a nice day :)