README.md

metadata

title: Model Test
emoji: 🔥
colorFrom: red
colorTo: purple
sdk: gradio
sdk_version: 5.34.1
python_version: 3.12.12
app_file: app.py
pinned: false
suggested_storage: large
models:
  - hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v
  - hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v
  - Wan-AI/Wan2.2-TI2V-5B-Diffusers
  - Wan-AI/Wan2.2-T2V-A14B
  - Qwen/Qwen3.6-35B-A3B
tags:
  - image-to-video
  - text-to-video
  - zero-gpu
  - hunyuanvideo-1.5
  - wan2.2
  - qwen
  - llm

Video Model ZeroGPU Test

This Space runs a Gradio test app for:

• hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v (text-to-video)
• hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v (image-to-video)
• Wan-AI/Wan2.2-TI2V-5B-Diffusers (text/image-to-video, 5B)
• Wan-AI/Wan2.2-T2V-A14B (text-to-video, 14B — higher quality, longer generation time)
• Qwen/Qwen3.6-35B-A3B (LLM — used for smart prompt enhancement)

Default settings:

• HunyuanVideo-1.5-T2V as default model
• Resolution: 848×480 (HunyuanVideo) / 1280×704 (Wan2.2)
• 81 frames
• 30 inference steps
• CFG 6.0

LLM Prompt Enhancement:

Enable the "Enhance prompt with LLM" checkbox to automatically expand your short prompt into a detailed, cinematic video description using Qwen3.6-35B-A3B. You can also click "Enhance now" to manually enhance the prompt before generation.

Usage:

Choose a model, upload a start image when needed (required for I2V), adjust the prompt, then click Generate.

By default, the app pre-downloads HunyuanVideo-1.5-T2V assets on startup with PRELOAD_MODELS=HunyuanVideo-1.5-T2V. You can set PRELOAD_MODELS=none to disable startup preloading and use the Download selected model button instead.

If the Space has persistent storage mounted at /data, downloads are stored there. Otherwise, they are stored in temporary Space storage and may need to be downloaded again after restart.