logoalt Hacker News

arcanemachineryesterday at 5:51 PM1 replyview on HN

Have you tried the '--jinja' flag in llama-server?


Replies

abhikul0yesterday at 6:32 PM

Yes, it fails too. I’m using the unsloth q4_km quant. Similarly fails with devstral2 small too, fixed that by using a similar template i found for it. Maybe it’s the quants that are broken, need to redownload I guess.