09-24-2026, 07:31 AM
Interesting, I hadn't tried that. The pattern around speculative decoding has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
|
Tensor cores actually helping here
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
Tensor cores actually helping here - by emberrunner - 09-24-2026, 03:44 AM
Re: Tensor cores actually helping here - by amberpatch25 - 09-24-2026, 04:55 AM
Re: Tensor cores actually helping here - by phaseport36 - 09-24-2026, 06:09 AM
Re: Tensor cores actually helping here - by orbitbench6 - 09-24-2026, 07:31 AM
Re: Tensor cores actually helping here - by llm_wren - 09-24-2026, 03:51 PM
Re: Tensor cores actually helping here - by amberpatch25 - 09-26-2026, 04:14 AM
|