Use different seeds to get different outputs

[miquel-espinosa]

Hi.
I would like to know if it is possible to use different seeds to get different outputs from terramind.
At the moment, I'm simply seeding before calling the model.

seed=1234
import random
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
if torch.cuda.is_available():
    torch.cuda.manual_seed_all(seed)

Is this the correct approach?
Thanks,

[blumenstiel]

torch.manual_seed(seed) + cuda might be enough, but I have not tried myself. Note that only the decoder generations are using noise. The token prediction of the TerraMind Decoder not.