Tensor Puzzles: Let's Play

[@Laplacey]

thank you soooooomuch! I learned a lot from this

[@larryabecid2819]

Thank you for helping me prepare for tensor interviews

[@danielcampos123]

For puzzle 8, you could just do: a @ triu(a.shape[0])

[@jcbmrshll]

great video!

[@skanderbegvictor6487]

For puzzle 13, I think we can use a zeros and then assign vector a to positions and then only broadcast till j

[@corgirun7892]

Very interesting

[@JohnSmith-he5xg]

Its pretty unclear what you're going for with the coloring scheme. Blue values are higher than yellow? Blue * yellow should be??? IMO gray scale would have been a better choice.

[@bubster2a]

Yellow are positive, blue are negative. More blue is a smaller number and more yellow is a bigger number

[@boussouarsari4482]

My way of doing compress! def compress(g: TT["i", bool], v: TT["i"], i:int) -> TT["i"]: small = v[g] i = g.shape[0] j = small.shape[0] return small @ eye(i)[:j,:]