Ghost In The Stack (@GhostInTh3Stack) header
Ghost In The Stack (@GhostInTh3Stack) profile picture

Local AI · Stretching a Single RTX 4090 24GB Form Over Function

Joined May 2024
Replying to @dimeiza
Sorry for the mostly Spanish on the screenshoot, but IQ3_S performs outstanding on my machine. You can also see the thought process in english. It even test the results where correct. RTX 4090 + 64GB RAM.
2
730
Wait what? We just had Qwen Image 2.1 and they already talking about 3.1? Please guys, gimme a break, i can't keep up 😭
2
205
Replying to @RobotCleopatra
So difficult to match the quality of your generations!
2
16
Direct comparison of Krea2 vs Qwen Image 2.1 Same prompt, same output resolution, same seed, no loras. Krea 8 steps CFG 1, Qwen 25 steps CFG 1 (Recommended values ​​for both). Judge the result for yourselves. Firts pic Krea2, second pic Qwen Image 2.1. Which one do you think wins? More pics ⤵️
3
3
339