Open Willmish opened 6 months ago
, Add quantisation flag!
NOTE and UPDATE on this issue. Since we are now using multiple-GPU, we may need to use torch.cuda.manual_seed_all
according to https://pytorch.org/docs/stable/generated/torch.cuda.manual_seed_all.html and https://pytorch.org/docs/stable/generated/torch.cuda.manual_seed.html#torch.cuda.manual_seed
well @Willmish will revisit
Guidelines: https://pytorch.org/docs/stable/notes/randomness.html
Inspiration from ADL GCW : https://github.com/Foalfloater66/c197-gcw/pull/20
Might need to set: https://pytorch.org/docs/stable/generated/torch.use_deterministic_algorithms.html#torch.use_deterministic_algorithms