Anima Twin Turbo - Public Edition

⚠️ USE THE UPLOADED WORKFLOW, THE IMAGES META-DATA IS A PROTOTYPE WORKFLOW.

🧪 The images include prompts.

📢 Here's the public release of my Anima Twin Turbo Workflow! This new version includes two new good generation seeds pre-fixed into the K-Samplers.

🖼️ Preferred image size is 1024x1536 or 1536x1024. Other sizes and resolutions work too!

🎨 This workflow is compatible and tested with all @circlestone_labs released Anima models and LoRAs, and I would highly recommend downloading, testing, and using all of their fine AI tools. I found the base and aesthetic 1.1 models to be the most stable in this generations from this workflow. My recommended combination is base model + aesthetic LoRA + RL LoRA on the first pass with base model + (optional turbo LoRA) + your desired style LoRAs on the second pass. This gives a great foundation for the original image generation and a stylized finishing pass to finalize the composition.

💭 It may also help to put your desired style LoRAs into both passes, as that can emphasize the style more, but it may overbake the generation (depending upon the LoRA and settings).

👀 This is a single model, dual-phase Anima workflow custom tuned and tested to enhance anime styling. It is best used with the base model. I recommend normal generation on the first past and turbo LoRA set up on the second pass. It includes Group Bypassing, Power LoRA Loaders, a Noise Buffer finish, and a Clean VRAM node. It uses commonly available and highly rated custom ComfyUI nodes from rgthree and ComfyUI Easy use. It includes recommended default generation settings and fixed generation nodes. The showcase video starting frames were made using the workflow, and the showcase images include several prototype workflows, LoRA character tests, LoRA style tests, simple prompts, and detailed prompts.