Thanks for your great work.
I want to use the inference of sparse conv operation, but the code doesn't provide such a function. Therefore, I am rewriting it like this.
When only_forward=True, the input will be processed directly without the operation of ctx.
However, I found there are nan using such a method. Do you have any ideas for solving it?
Thanks for your great work. I want to use the inference of sparse conv operation, but the code doesn't provide such a function. Therefore, I am rewriting it like this. When only_forward=True, the input will be processed directly without the operation of ctx. However, I found there are nan using such a method. Do you have any ideas for solving it?
class ConvolutionFunction(Function):