Comment by forgotTheLast
12 hours ago
Zero temp just uses argmax, which is what softmax approaches if you take the limit of T to zero anyway. So it could very well be deterministic.
12 hours ago
Zero temp just uses argmax, which is what softmax approaches if you take the limit of T to zero anyway. So it could very well be deterministic.
No comments yet
Contribute on Hacker News ↗