The Lottery Ticket Hypothesis: Applying the Mask
Beginner
The problem, from first principles
Once a mask is chosen, the sparse network is just the original weights with the pruned entries forced to zero. The paper's key step is to reset the kept weights to their initial values, which this question does not do yet.
From theory to code
Implement apply_mask(w, mask), which zeros the weights where the mask is False.
Constraints
maskis boolean and matches the shape ofw.
Hints
Hint 1
Multiply element-wise, letting False become 0.
Keywords
research-papersclassical-mlpruninglottery-ticket
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.