search

Zhendong-Wang / Diffusion-Policies-for-Offline-RL

Apache License 2.0
219 stars 33 forks source link

Suggestion to pre-reload the data into GPU memory #2

Open ChenDRAG opened 1 year ago

ChenDRAG commented 1 year ago

By preloading the training data into the GPU memory first, we can save about 95% of CPU usage (at least on my PC).