ilgu kim
ilgu kim
Good day ! I think chapter 8 is calling me out from there Is there chance for me?
as people suddenly united for the translation, I think translator should have a chat for the translation. there must be some translation rule, or convention. any idea or channel? import...
I`m trying to copy this tutorial in tpu-vm-base version cloud-tpu and it works well for training. I was trying to run the code from tensorflow-tpu-vm,, maybe it triggered the bug
회사 코드와 이름 받아오는 코드로 직접 구성하셔서 일단 사용하시면 어떨까요
@lewtun Hi there! I just checked that My PR worked fine! maybe you could fill the chapter 8 up there!
thanks for your help!
I tested a simple benchmark using PyTorch and Rust’s candle. I performed matmul, increasing the number of repetitions on a tensor of size (1000,1000) in float32 format. torch n=99, 0.183099s...
@edgarriba I have reviewed the issue you posted. The test I conducted was a simple repetition of the matmul operation in a for loop with the 1000x1000 matrix mentioned in...
@edgarriba don`t know if this what you wanted 3x3 tensor, ten milion times matmul result torch.mm, cpu 55.3527s torch.matmul, cpu, 55.5209s torch.mm, gpu(a6000), 112.1874s numpy.dot, cpu, 14.3464s numba jit +...