Loading...

  1. v7.2.0iOS
    - Support for Latent Consistency Models
    - Small fixes and improvements
  2. v6.4.0iOS
    - Improved support for Stable Diffusion XL ControlNet
    - Fixes and improvements
  3. v6.2.0iOS
    - Fast progress, now you can choose to either show a fast version of the current latent or decode it to get an improved version at the cost of slower diffusion
  4. v6.1.1iOS
    - Shortcuts support
    - Stable Diffusion XL
    - Inpainting support for every model
    - Improvements and bug fixes
    - Denoised progress
    - Textual inversion
    - Prompt history
    - Token count
  5. v6.1.0iOS
    - Shortcuts support
    - Stable Diffusion XL
    - Inpainting support for every model
    - Improvements and bug fixes
    - Denoised progress
    - Textual inversion
  6. v5.0.0iOS
    - Loading on demand for models, it's no longer necessary to manually load and unload the models
    - Memory and performance improvements
  7. v4.1.0iOS
    - Performance improvements
    - Fixes in img2img with different scale ratios
  8. v4.0.0iOS
    - Added new Samplers: DDIM, Euler, EulerA, DPM++ 2M Karras
    - Added support for Stable Diffusion 2 based models, no more black screens! This does require converting the models with the new Guernika Model Converter 4.0.0
    - Improve alignment between Guernika's output and python implementation output.
    - Some bugs fixes and performance improvements
  9. v2.1iOS
    - Prompt weighting, tokens between () are taken more into account and tokens between [] will be taken less into account.
    - Initial support for Textual Inversion embeddings, embeddings can be added when converting the TextEncoder.
    - Added token count.
    - Other fixes.
  10. v2.0iOS
    - Add support for newer models
    - General improvements
  11. v1.9.1iOS
    - Fix symbolic links, newly added symbolic links should work after app relaunch
    - Improvements on model downloading
  12. v1.9iOS
    - Allow creating links to models on external storage
    - Fix problems downloading models
  13. v1.8iOS
    - Improve model downloading
    - Keep screen on while performing long task
  14. v1.6iOS
    - Add initial support for Instruct Pix2Pix model, generate image variation using instruction ("make it look like a gold statue")
    - Fixes
  15. v1.4iOS
    - Improve support for different output sizes
    - Allow choosing compute units
  16. v1.3iOS
    - Added DPMSolver sampler
    - Allow disabling progress images
    - General fixes and improvements

Disclosure: Independent intel to help mobile builders succeed.

AI-powered analysis with editorial review, built from publicly available sources. Marlvel.ai is not affiliated with, endorsed by, or sponsored by Guernika, its developer, the app publisher, Apple, or Google Play. All trademarks, logos, and screenshots referenced remain the property of their respective owners.

Data licensed under CC-BY-NC 4.0