HunyuanVideo-1.5-480p-Turbo

This is a Turbo version of Tencent's HunyuanVideo-1.5, containing the 480p T2V transformer along with a specialized LoRA that enables ultra-fast 4-step inference. This optimized package offers the following advantages:

• Ultra-Fast: Video generation now requires only 4 steps instead of 50, leading to ~12x speed-up
• High-Quality: Maintains generation quality equivalent to the base model
• Efficient: Reduced inference time without sacrificing visual fidelity
• LoRA-Based: Uses a lightweight LoRA adapter for fast distillation

Features

• 4-Step Inference: Dramatically reduced sampling steps (50 → 4)
• Equivalent Quality: Delivers visuals on par with the 50-step base model
• Optimized Package: Contains 480p T2V transformer + Turbo LoRA weights
• Production-Ready: Ideal for applications requiring fast video generation
• Based on State-of-the-Art: Built on Tencent's advanced video generation technology

Usage

This model is planned to work with Aquiles-Image, providing an OpenAI-compatible API for video generation:

pip install aquiles-image
aquiles-image serve --model "hunyuanVideo-1.5-480p-turbo"

Learn more in the full documentation.

Example of a Video Generated with This Model

Generated with prompt: A direct continuation of the existing shot of a chameleon crawling slowly along a mossy branch. Begin with the chameleon already mid-step, camera tracking right at the same close, eye-level angle. After three seconds, its eyes swivel independently, one pausing to glance toward the lens before it resumes moving forward. Maintain the 100 mm anamorphic lens with shallow depth of field, dappled rainforest light, faint humidity haze, and subtle film grain. The moss texture and background greenery should remain consistent, with the chameleon's deliberate gait flowing naturally as if no cut occurred.

Acknowledgments

We extend our gratitude to:

• The Tencent Hunyuan Team for developing and open-sourcing HunyuanVideo-1.5
• The lightx2v team (@lightx2v) for training and providing the high-quality Turbo LoRA weights that enable 4-step inference

This repackaged version is made possible by their excellent work and commitment to the open-source community.

For more information about the original model, please visit the official HunyuanVideo-1.5 repository.