ComfyUI Node: Hunyuan Instruct Loader
Run ComfyUI workflows without the setup
No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.
Category
Hunyuan/Instruct
Inputs
model_name
- HunyuanImage-3.0-Instruct-Distil-INT8
- HunyuanImage-3.0-Instruct-Distil-NF4
force_reload BOOLEAN
attention_impl
- sdpa
- flash_attention_2
moe_impl
- eager
- flashinfer
vram_reserve_gb FLOAT
blocks_to_swap INT
moe_drop_tokens BOOLEAN
vae_dtype
- bfloat16
- float32
Outputs
HUNYUAN_INSTRUCT_MODEL
Extension: Comfy_HunyuanImage3
Professional ComfyUI custom nodes for Tencent's HunyuanImage-3.0 80B multimodal model, with multiple loading modes, smart memory management, large image generation with CPU offload, and optional prompt enhancement via LLM APIs.
Authored by EricRollei
Looking for a different node?
More nodes in Comfy_HunyuanImage3
- Hunyuan Cache Status V2
- Hunyuan Emergency Cleanup
- Hunyuan Empty Latent
- Hunyuan Generate with Latent
- Hunyuan 3 Clear Downstream Models
- Hunyuan 3 Loader (Multi-GPU BF16)
- Hunyuan 3 Force Unload (Nuclear)
- Hunyuan 3 Loader (Full BF16 GPU)
- Hunyuan 3 Loader (Full BF16)
- Hunyuan 3 Generate
- Hunyuan 3 Generate (HighRes Efficient)
- Hunyuan 3 Generate (Large/Offload)
Other Hunyuan/Instruct nodes
Run ComfyUI workflows without the setup
No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.