Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

🤩 [User Study] Report your bad examples! #19

Open
PommesPeter opened this issue May 12, 2024 · 6 comments
Open

🤩 [User Study] Report your bad examples! #19

PommesPeter opened this issue May 12, 2024 · 6 comments

Comments

@PommesPeter
Copy link
Contributor

PommesPeter commented May 12, 2024

Hi, all

Thank you for being so interested in Lumina-T2X. If you encounter images of poor quality, please feel free to report them in this issue to assist in our model's enhancement.

You can directly add a comment on this issue and use the template below:

prompt: <copy-paste prompt here>
image: <copy-paste image here>

You can also optionally report the hyperparameters you have used if it does influence the quality.

@PommesPeter PommesPeter changed the title 🤩 Seeking Feedback on Low-Quality Images 🤩 [Feedback] Welcome to share your bad examples! May 12, 2024
@PommesPeter PommesPeter pinned this issue May 12, 2024
@PommesPeter PommesPeter changed the title 🤩 [Feedback] Welcome to share your bad examples! 🤩 [User Study] Welcome to report your bad examples! May 12, 2024
@PommesPeter PommesPeter changed the title 🤩 [User Study] Welcome to report your bad examples! 🤩 [User Study] Report your bad examples! May 12, 2024
@jrd-rocks
Copy link

Lumina-Next-SFT

prompt: a giant mech reaching out to a skyscraper
image: image-1

prompt: dramatic digital painting of an alluring mermaid sitting on top of a rock that emerges from the wild and roaring sea, under the waterline the base of the rock and seafloor is littered with human skulls and bones, a deceptively tranquil and peaceful atmosphere
image: image

Maybe prompts like these aren't fair, but i find that, apart from the big hosted-only models, generative image models generally just can't do slightly more complex scenes like this. Describing how elements look often goes fine, but describing the composition of (objects on) a scene, often goes awry unless it's really basic.

(The gens aren't very coherent (mech head and limbs, mermaid face and tail), but that's not why i post them, I'm sure further training will help with that)

@gaopengpjlab
Copy link
Contributor

@jrd-rocks Your feedback is greatly appreciated.

@heart-du
Copy link

为什么绿植感觉一直生成的不太好。
Uploading image.png…

@heart-du
Copy link

A garden where the shrubs are three-dimensional and layered, and every leaf and branch can be seen.
不够立体和真实

@gaopengpjlab
Copy link
Contributor

Thanks for your feedback. We will look into the problem.

@gaopengpjlab
Copy link
Contributor

gaopengpjlab commented Jun 28, 2024

Can you share your failure cases?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants