is there a reason why you prefer:
rand_choices=samples[torch.randint(len(samples),(7,))] #’'7 choices ‘’
vs
**perm = torch.randperm(tensor.size(0))**
?
Just curious
is there a reason why you prefer:
rand_choices=samples[torch.randint(len(samples),(7,))] #’'7 choices ‘’
vs
**perm = torch.randperm(tensor.size(0))**
?
Just curious