HunyuanVideo1.5

Quick Start

1. Prepare docker environment:

docker pull lightx2v/lightx2v:25111101-cu128

2. Run the container:

docker run --gpus all -itd --ipc=host --name [container_name] -v [mount_settings] --entrypoint /bin/bash [image_id]

3. Prepare the models

Please follow the instructions in HunyuanVideo1.5 Github to download and place the model files.

4. Running

Running using bash script

# enter the docker container

git clone https://github.com/ModelTC/LightX2V.git
cd LightX2V/scripts/hunyuan_video_15

# set LightX2V path and model path in the script
bash run_hy15_t2v_480p.sh

Running using Python code: edit the paths in the selected file and run it from the repository root after installation.

• Text to video with LightTAE
• Image to video with quantization
• Distilled text to video

See the Python API guide for configuration and generator lifecycle.

5. Check results

You can find the generated video files in the save_results folder.

6. Modify detailed configurations

You can refer to the config file pointed to by --config_json in the script and modify its parameters as needed.