I created a new quantum technique using pytorch!
So I’m trying to see how well this technique works on my server computer, Raspberry Pie B3+ and the Jetson TX2.
But there’s a problem.
I don’t know how to change 32bit operation to 4bit operation.
Now the value itself is quantified through clipping, but it still operates at 32bit in GPU or CPU.
I’ve been looking for ways to use CUDA, but I haven’t found a clue.
I really don’t know where to approach this to solve.
How can I change a 32-bit operating system from CPU and GPU to a 4-bit or 8-bit operating system?
I’d really appreciate it if you let me know.