still bad compare to hunyuan dit try this complex violinist prompt, one shot no cherry picking

[s9anus98a]

kittychan is a beautiful female acoustic violin player with long brown hair which holds a violin, in the style of romantic riverscapes, yuumei, gil elvgren, hannah flowers, uhd image, album covers, synthetism-inspired ,ultra detailed realistic face

[gaopengpjlab]

Thanks for your timely feedback. We will continue to improve the foundational ability of our model. Currently, you can use extrapolation at 1536 to slightly improve the performance.

[gaopengpjlab]

In the technical report, hunyuan dit claim to use more than 1 billion image-text pairs for training while Lumina-Next utilize 20 million image-text pairs. We will continue to expand the versatility and domains of our training set.

[s9anus98a]

In the technical report, hunyuan dit claim to use more than 1 billion image-text pairs for training while Lumina-Next utilize 20 million image-text pairs. We will continue to expand the versatility and domains of our training set.

ooh now i see thanks for your reply. you & team doing great job. it just the parameter billion Vs million. i will help with testing since i don't know how to code lmao

[gaopengpjlab]

You are welcome to provide more feedback. We will carefully take your feedback into consideration and design efficient strategy to solve them.