When I use this model the result I get is some patches of color
I use webui to generate images. I using vae of sdxl, but finally I get picture like this:
I have been using sd1.5 before and don't particularly understand what needs to be done to use sdxl.
I observed from the image metadata provided in this discussion that you are still utilizing Negative Textual Inversion, such as [(AS-YoungV2-neg:1.1), (BadDream:1.6), (badhandv4:1.4), (BadNegAnatomyV1-neg:1.3), (EasyNegative:1.4), (FastNegativeV2:1.1)]. It is noteworthy that these negative embeddings can potentially impact the outcomes of generated images using the AnimagineXL 3.0 model. I suggest utilizing the Negative prompt that has been provided in our recommended settings, as outlined in recommended-settings. Then, demonstrate the results to us again, especially if the generated images remain consistent with the ones you initially provided.
I observed from the image metadata provided in this discussion that you are still utilizing Negative Textual Inversion, such as [(AS-YoungV2-neg:1.1), (BadDream:1.6), (badhandv4:1.4), (BadNegAnatomyV1-neg:1.3), (EasyNegative:1.4), (FastNegativeV2:1.1)]. It is noteworthy that these negative embeddings can potentially impact the outcomes of generated images using the AnimagineXL 3.0 model. I recommend refraining from employing Textual Inversion in order to optimize the results.
I think this has little to do with Negative Textual Inversion. When I removed all the negative cue words, the picture still looked like this color block.
Here my result
metadata my image using same prompt
1girl,c.c.,code geass,white shirt,long sleeves,turtleneck,sitting,looking at viewer,eating,pizza,plate,fork,knife,table,chair,table,restaurant,cinematic angle,cinematic lighting,masterpiece,best quality,
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 2783292315, Size: 1024x1024, Model hash: 1449e5b0b9, Model: animagine-xl-3.0, CFG Rescale phi: 0, Version: v1.7.0
It seems there might be an issue with the AnimagineXL 3.0 model you downloaded, as the generated images consistently exhibit artifacts. Please attempt to download the model directly from HuggingFace to address this concern.
Here my result
metadata my image using same prompt 1girl,c.c.,code geass,white shirt,long sleeves,turtleneck,sitting,looking at viewer,eating,pizza,plate,fork,knife,table,chair,table,restaurant,cinematic angle,cinematic lighting,masterpiece,best quality, Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 2783292315, Size: 1024x1024, Model hash: 1449e5b0b9, Model: animagine-xl-3.0, CFG Rescale phi: 0, Version: v1.7.0It seems there might be an issue with the AnimagineXL 3.0 model you downloaded, as the generated images consistently exhibit artifacts. Please attempt to download the model directly from HuggingFace to address this concern.
I really didn't think it was a download problem. I checked the sha256 value of the file I originally downloaded and found that it was inconsistent. After re-downloading the model, it can run normally. Thanks.