ComfyUI Node
Vision Patch Encode
A ComfyUI node in ✨ LF Nodes/Analytics with 4 inputs and 1 output.
Vision Patch Encode
- clip_vision
- image
- ui_widget
- clip_vision_output
◄crop▾►
Category✨ LF Nodes/Analytics
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| clip_vision | CLIP_VISION | Model from core CLIPVisionLoader, e.g. dino_v3_vit_h.safetensors. | |
| image | IMAGE | Image or batch to encode. | |
| crop | COMBO | As core CLIPVisionEncode: center crops the square, none stretches the full image. | |
| ui_widgetopt | LF_COMPARE | [object Object] | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| clip_vision_output | CLIP_VISION_OUTPUT | Core CLIP_VISION_OUTPUT with patch tokens (last_hidden_state), image_embeds and image_sizes. |