ComfyUI Node
Eric Qwen-Image Load Model
A ComfyUI node in Eric Qwen-Image with 8 inputs and 1 output.
Eric Qwen-Image Load Model
- pipeline
◄model_path►
◄precisionbf16►
◄devicecuda►
◄keep_in_vramtrue►
◄offload_vaefalse►
◄attention_slicingfalse►
◄sequential_offloadfalse►
◄vae_tile_blendcosine►
CategoryEric Qwen-Image
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| model_path | STRING | Path to Qwen-Image or Qwen-Image-2512 model directory | |
| precisionopt | COMBO | bf16 | Model precision (bf16 recommended for RTX 40/50 series) |
| deviceopt | COMBO | cuda | Device to load model on |
| keep_in_vramopt | BOOLEAN | true | Cache model between runs (faster but uses VRAM) |
| offload_vaeopt | BOOLEAN | false | Move VAE to CPU during transformer inference (saves ~1 GB) |
| attention_slicingopt | BOOLEAN | false | Trade speed for lower peak VRAM |
| sequential_offloadopt | BOOLEAN | false | Extreme VRAM savings via sequential CPU offload |
| vae_tile_blendopt | COMBO | cosine | Tile-seam blending for tiled VAE decode/encode. • cosine — C¹-smooth blend (zero slope at tile edges), eliminates faint grid lines on smooth gradients. • linear — original diffusers behaviour. Slightly sharper at tile edges; may show faint seams at high resolutions. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| pipeline | QWEN_IMAGE_PIPELINE | — |