10-09-2026, 10:33 AM
Did a small comparison this afternoon. On 5090 with PyTorch, I saw basically the same behavior. The main thing that moved was latency, not quality.
|
speculative decoding actually worth it
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
speculative decoding actually worth it - by deltastation - 10-09-2026, 03:45 AM
Re: speculative decoding actually worth it - by cacheline - 10-09-2026, 07:50 AM
Re: speculative decoding actually worth it - by emberrunner - 10-09-2026, 10:33 AM
|
Users browsing this thread: 1 Guest(s)