19 lines
775 B
Markdown
19 lines
775 B
Markdown
# LLMStudioPictureDescribe
|
||
Send a single image tensor to LM Studio and receive a richly detailed caption using the selected VL model.
|
||
|
||
## Inputs
|
||
- `base_url`, `model_name` – LM Studio endpoint + model.
|
||
- `image` – ComfyUI IMAGE tensor to caption.
|
||
- `caption_prompt` – instructions for what to focus on.
|
||
- `temperature`, `verbose` – language style + logging.
|
||
|
||
## Outputs
|
||
- `caption` – generated paragraph describing the input image.
|
||
|
||
## Usage Tips
|
||
1. Insert the node after any IMAGE output (e.g., generated frames, uploads).
|
||
2. Tweak the prompt to emphasize facial features, outfits, etc., then route the caption to ShowText or Control Panel.
|
||
|
||
## Additional Resources
|
||
- [Full documentation](docs/nodes/llm-studio-describe/LLM_STUDIO_VIDEO_DESCRIBE.md#related-nodes)
|