Fascinatio.Redmond is here!
I'm grateful for the GPU time from Redmond.AI that allowed me to finish this model!
This is a generalist model fine-tuned on Pixart-Alpha.
The model has a high capacity to generate realistic, artistic images, cars, people, and a wide variety of themes. It's a versatile model. Its a little improvement from the promising model Pixart-Alpha.
This model has been trained for approximately 60k steps.
I really hope you like the model and use it.
If you like the model and think it's worth it, you can make a donation to my Patreon or Ko-fi.
Follow me in my twitter to know before all about new models:
Description
FAQ
Comments (15)
Is this 1.5 or sdxl?
It's neither - It's a different model w/ a different architecture: https://github.com/PixArt-alpha/PixArt-alpha
1.5 VAE but does 1280x1024
@DSlater Thank you!
I try to understand, why this model make the exact same picture as the last model loaded and can't be load on A1111 startup ? Maybe i do smth wrong.
Usually when this happens it means it hasn't been loaded. I've seen that in the past, with models or VAEs.
I'm struggling right now with this and ComfyUI, and haven't seen auto1111 mentioned anywhere, maybe it's not available yet
It will not load in A1111 until somebody does an extension for it. It is a completely different architecture than stable diffusion. Your best bet is comfyui with manager installed. You can use this workflow: https://github.com/city96/ComfyUI_ExtraModels/files/13617463/PixArtV3.json and use the manager to get the missing extensions for it. It will require some leg work to get the related models t5v.1.1 and the pixart model in the proper folder
Please provide some technical information to run it. Can this be run on Automatic1111? The VAE, ram, vram requirements, etc. Thank you for your contribution to the model.
1.5 vae, I use ema-560000-ema-pruned satisfactorily. I'm using ComfyUI --lowvram (8Gb RTX 2070)
Can this be used with Comfy UI? What extensions are necessary?
@lawrence_cjs Thank you!
I'm trying to understand what this model is. The file sizes make no sense for a Pixart Alpha finetune.
I suppose it's only the UNET part, as far as I remember the text encoder is the biggest part of the model and is not supposed to be put in safetensors or ckpt
It would be good to add some example generations with prompts to show how to use the model.