Closed aiyzs closed 10 months ago
Same problem
Couple things you can try that seem to have no/ little performance impact: (1) Batch size 6 -> 4 for the first training stage (2) n_neurons for the proposal network 32 -> 16 (3) (Last resort) reduce number samples for either the proposal or nerf network Let me know if some combination of these steps can resolve the issues.
I want to run the model on my 3090 device, but there is not enough memory. How can I adjust the relevant parameters to make it work