Looking for additional instructions
pengzhangzhi opened this issue · comments
A sincerely Thank you to srush for such an amazing exercises. I have questions when dealing with the following puzzles. I am wondering if there are more cues to support me complete them.
## Puzzle 2 - sum
Compute [sum](https://numpy.org/doc/stable/reference/generated/numpy.sum.html) - the sum of a vector.
## Puzzle 7 - cumsum
Compute [cumsum](https://numpy.org/doc/stable/reference/generated/numpy.cumsum.html) - the cumulative sum.
## Puzzle 12 - compress
Compute [compress](https://numpy.org/doc/stable/reference/generated/numpy.compress.html) - keep only masked entries (left-aligned).
def compress(g: TT["i", bool], v: TT["i"], i:int) -> TT["i"]:
return torch.tensor([v[j] if g[i] else None for j in range(i)])
## Puzzle 15 - bincount
Compute [bincount](https://numpy.org/doc/stable/reference/generated/numpy.bincount.html) - count number of times an entry was seen.
For Puzzle 2 and 7 my recommendation would be to try to come up with a special vector or matrix that can compute the necessary output values. Then it just becomes a matrix multiplication.
For puzzle 15, I would think about one-hot vectors and how they can be used.
Thanks. your instructions are very insightful!