[2.1.864] New model list #2067
Replies: 7 comments 6 replies
-
Hi, thank you for the great work. Do you have any suggestions on how to fix an issue where my run.bat file is no longer reading information from the config.txt file after the 2.1.864 update? |
Beta Was this translation helpful? Give feedback.
-
Great! I have a request though. Can you share any prompt tips to get quality results? I am used to those insane long prompts From A1111. |
Beta Was this translation helpful? Give feedback.
-
(To always download new models, use --always-download-new-model ) |
Beta Was this translation helpful? Give feedback.
-
Sir, why everytime gradio.live runs. |
Beta Was this translation helpful? Give feedback.
-
Hey can anyone suggest how to train this realistic model realisticStockPhoto_v20 with my own photos locally? |
Beta Was this translation helpful? Give feedback.
-
Hi, can I get more realistic anime style results with v100 or should I look for other model? |
Beta Was this translation helpful? Give feedback.
-
no good ART model for Fooocus? |
Beta Was this translation helpful? Give feedback.
-
Below model list will be used by Fooocus since 2.1.864 as default:
If you install Fooocus after 2024 Jan 27, you do not need to do anything and all downloads are automatic.
To make the software user-friendly, if you already have previous versions of JuggernautXL or realisticStockPhoto, Fooocus will not download new models.
To always download new models, use
--always-download-new-model
The JuggernautXL and RealisticStockPhoto are just version difference and they should be similar to before
The anime model is animaPencilXL_v100. Note that:
animaPencilXL_v100 = animagine-xl-3.0 + blue_pencil-XL
blue_pencil-XL = a merge of about 150 sdxl models (including Kohaku-XL beta)
In the final round of test, the model animaPencilXL_v100 and AAMXL(another animagine v3 mix) are both slightly more flexiable to use than animagine-xl-3.0, especially when the prompt is short or casually written. (the original animagine-xl would require formatted prompts, and shorter prompts or casual prompts sometimes will cause artifacts or saturation problems. Using a merge from animagine-xl, like animaPencilXL_v100 or AAMXL, can fix this problem.)
Sanity Check
Anime Preset, Default outputs with same seed 12345
1girl, garden
1girl, sitting on grass
1girl, fire magic
1girl, white wedding dress
1girl, armor
1girl, dark magic
1girl, dragon
Beta Was this translation helpful? Give feedback.
All reactions