3 ms·
You could also preload your datasets to GPU, if its RAM is large enough.
by lostmsu 5y ago
You could also preload your datasets to GPU, if its RAM is large enough.
- celrod 5y agoThe Flux.jl example did this. A PR to the PyTorch example to do this would be welcome: https://github.com/chriselrod/LeNetTorch https://github.com/chriselrod/LeNetTorch