tm-hf-repo/seedream-pixar

seedream pixar

Model description

This is a 3000 step flux kontext lora trained on the seedream prompt:

convert this to a 3d pixar style, maintain the same pose, colors, skin tone, and hair. beautiful 3d rendered hair texture, empty white background

Trigger words

You should use tmc_pix to trigger the image generation.

Recommendations:

• Try increasing the steps to 40 during generation and guidance scale to 3-3.5
• Highly recommend adding "maintain the same skin tone" to the "tmc_pix" prompt
• Note due to the training data, it pretty much always skews towards smiling, seedream pixar seems to overfit to smiling characters

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.

Training at fal.ai

Training was done using fal.ai/models/fal-ai/flux-kontext-trainer.

INPUT IMAGE

Base output at seed=42 and prompt="tmc_pix" everything else defaults

Base output at seed=42, guidance=3.5, prompt="tmc-pix, maintain the same skin tone",