parser.add_argument('--dropout',
type = float, default = 0.1)
-parser.add_argument('--synthesis_sampling',
- action='store_true', default = True)
+parser.add_argument('--deterministic_synthesis',
+ action='store_true', default = False)
parser.add_argument('--no_checkpoint',
action='store_true', default = False)
for s in range(first, input.size(1)):
output = model(input)
logits = output[:, s]
- if args.synthesis_sampling:
+ if args.deterministic_synthesis:
+ t_next = logits.argmax(1)
+ else:
dist = torch.distributions.categorical.Categorical(logits = logits)
t_next = dist.sample()
- else:
- t_next = logits.argmax(1)
input[:, s] = t_next
return results