10-10-2026, 07:17 AM
A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at prefill/decode split. Not glamorous advice, but repeatable baselines save a lot of time.
|
small win thread
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
small win thread - by quiettoken - 10-09-2026, 05:00 AM
Re: small win thread - by quiettoken - 10-10-2026, 05:24 AM
Re: small win thread - by rocmrunner40 - 10-10-2026, 06:21 AM
Re: small win thread - by cacheplane98 - 10-10-2026, 06:28 AM
Re: small win thread - by thetagrid15 - 10-10-2026, 07:17 AM
Re: small win thread - by rocmrunner40 - 10-10-2026, 08:08 AM
Re: small win thread - by relaysupply90 - 10-10-2026, 10:21 AM
Re: small win thread - by rocmplane - 10-10-2026, 11:31 AM
Re: small win thread - by sparrowgrid - 10-10-2026, 04:09 PM
Re: small win thread - by vectorshed43 - 10-10-2026, 05:33 PM
|