🚀 Run Hunyuan Video 1.x with ComfyUI with provisioning — RunPod.io Deployment

Pod running on L40S

A streamlined and automated environment for running ComfyUI with Hunyuanvideo models, optimized for use on RunPod.io.

🔧 Features

Automatic model and LoRA downloads via environment variables.
Supports advanced workflows for video generation and enhancement using pre-installed custom nodes.
Compatible with high-performance NVIDIA GPUs

Variable	Description
`COMFYUI_EXTRA_ARGUMENTS`	Additional arguments for ComfyUI CLI

Model Type	Model	Safetensors/GGUF
Diffusion Model	`HF_MODEL_DIFFUSION_MODELS[1-20]`	`HF_MODEL_DIFFUSION_MODELS_FILENAME[1-20]`
Checkpoints	`HF_MODEL_CHECKPOINTS[1-20]`	`HF_MODEL_CHECKPOINTS_FILENAME[1-20]`
Text Encoders	`HF_MODEL_TEXT_ENCODERS[1-20]`	`HF_MODEL_TEXT_ENCODERS_FILENAME[1-20]`
Clip Vision	`HF_MODEL_CLIP_VISION[1-20]`	`HF_MODEL_CLIP_VISION_FILENAME[1-20]`
Audio Encoders	`HF_MODEL_AUDIO_ENCODERS[1-20]`	`HF_MODEL_AUDIO_ENCODERS_FILENAME[1-20]`
Model patches	`HF_MODEL_PATCHES[1-20]`	`HF_MODEL_PATCHES_FILENAME[1-20]`
VAE	`HF_MODEL_VAE[1-20]`	`HF_MODEL_VAE_FILENAME[1-20]`
Upscalers	`HF_MODEL_UPSCALER[1-20]`	`HF_MODEL_UPSCALER_PTH[1-20]`
Loras	`HF_MODEL_LORA[1-20]`	`HF_MODEL_LORA_FILENAME[1-20]`
VLM/mmproj	`HF_MODEL_VL[1-20]`	`HF_MODEL_VL_FILENAME[1-20]`
SAM segmentation	`HF_MODEL_SAMS[1-20]`	`HF_MODEL_SAMS_FILENAME[1-20]`
Latent Upscale	`HF_MODEL_LATENT_UPSCALE[1-20]`	`HF_MODEL_LATENT_UPSCALE_FILENAME[1-20]`

Type	Model	Safetensors/GGUF	/workspace/ComfyUI/
File	`HF_MODEL[1-20]`	`HF_MODEL_FILENAME[1-20]`	`HF_MODEL_DIR[1-20]`
Dir	`HF_FULL_MODEL[1-20]`		`HF_FULL_MODEL_DIR[1-20]`

Variable	Description
`CIVITAI_MODEL_LORA_URL[1-50]`	Direct download link for LoRAs

Variable	Description
`WORKFLOW[1-50]`	download link (compressed or plain)

Package	Version
flash_attn	`2.8.3`
sageattention	`2.2.0`

Processor	Compute Capability	SM
A40	`8.6`	`sm_86`
L40S	`8.9`	`sm_89`