Stable Diffusion · LoRA

Spontaneous Combustion LoRA for Wan I2V 14B

LoRA for Wan2.2 I2V 14B that generates dramatic spontaneous combustion fire effects in image-to-video generation, trained on the trigger phrase 'Spontaneous…

About this model

This LoRA specializes in creating dramatic spontaneous combustion effects from static images. It transforms ordinary objects and scenes into dynamic sequences of sudden, mysterious ignition with realistic fire propagation, smoke, and ember effects. The model is designed for the Wan-AI Wan2.2-I2V-A14B image-to-video pipeline, enabling creators to animate scenes where objects burst into flames without any external trigger.

The LoRA was trained on the trigger phrase "Spontaneous Combustion" and is distributed as two checkpoint files: a low-noise variant and a high-noise variant, both in SafeTensors format. The combined file size is 73.22 MB. To use it, load the base Wan I2V model in ComfyUI, add a Load LoRA node, and select either the low-noise or high-noise checkpoint. Connect the LoRA output to the KSampler and set the strength_model to 0.8-1.0 as a starting point.

The project was made possible with sponsorship and GPU time from Redmond Ai. It is released under the permissive CC0-1.0 license, allowing free commercial and non-commercial use without restrictions. On Civitai, the model has received 51 likes and over 1,000 downloads, indicating active adoption among Wan I2V users.