ComfyUI Node
Qwen3 VL Caption (Inverse Prompt)
A ComfyUI node in image/caption with 10 inputs and 1 output.
Qwen3 VL Caption (Inverse Prompt)
- image
- text
◄model_path▾►
◄dtypeauto►
◄keep_model_loadedfalse►
◄unload_other_modelstrue►
◄lang中文►
◄seed1►
◄max_side512►
◄video_fps16.0►
◄instruction—►
Categoryimage/caption
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| model_path | COMBO | 0 options: | |
| dtype | COMBO | auto | 3 options: auto, 4bit, 8bit |
| keep_model_loaded | BOOLEAN | false | — |
| unload_other_models | BOOLEAN | true | — |
| lang | COMBO | 中文 | 3 options: 中文, English, bbox |
| seed | INT | 10–4294967295 | — |
| max_side | INT | 512256–2240 | — |
| imageopt | IMAGE | — | |
| video_fpsopt | FLOAT | 16.01–200 | — |
| instructionopt | STRING | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| text | STRING | — |