4 ms·
60 to 70 on a 5080, but only tinkering for now. The smaller models seem exceptionally good for what they are, and some can even do OCR reliably.
by politelemon 7mo ago
60 to 70 on a 5080, but only tinkering for now. The smaller models seem exceptionally good for what they are, and some can even do OCR reliably.
- cpburns2009 7mo agoWhat quantization are you running on the 5080? I'm waiting to receive mine.