SpectraSynth_SDXL_TCD

SpectraSynth_SDXL_TCD is unlike any other SDXL model, made for fast image generation and uses a minimum of only 2 steps with a CFG scale of 1.3, without loosing image clarity. It has the TCD (Trajectory Consistency Distillation) Lora embedded.

Sample Images:

All sample images are generated on a ASUS ROG laptop with an GTX 1070 GPU.

Q&A

Q: Can I use a Lightning Lora together with the model?

A: No as that will result in images being burned as the model is already optimized for 2 STEP image generation,

The default ComfyUI Workflow I use for my SpectraSynth and Azazeal Anime Diffusion models is attached to this article, it also include the prompt used. Feel free to add upscale nodes, etc to the workflow as per your needs.

You can download my Azazeal Anime Diffusion model on TensorArt