TCN: Causal Padding
Beginner
The problem, from first principles
A causal convolution must not look at future samples. Padding on the left by the dilated kernel span keeps the output length equal to the input and keeps the output causal.
From theory to code
Implement causal_padding(kernel, dilation), the left padding for a causal dilated convolution.
Constraints
- Pad on the left only.
Hints
Hint 1
Multiply kernel minus one by the dilation.
Keywords
research-paperssequence-modelsconvolutionsequence
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.