ComfyUI Node
LUNA VLM Chat
A ComfyUI node in LUNA/VLM with 9 inputs and 1 output.
LUNA VLM Chat
- llm_model
- image
- text
◄promptDescribe this image in detail for use as an image generation prompt.►
◄system_promptYou are a helpful assistant that describes images concisely and accurately for use in AI image generation prompts.►
◄max_tokens256►
◄temperature0.70►
◄repetition_penalty1.15►
◄prepend_text►
◄append_text►
CategoryLUNA/VLM
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| llm_model | LLM_MODEL | From LUNA VLM Loader (must have mmproj) | |
| image | IMAGE | Input image for the VLM to describe | |
| prompt | STRING | Describe this image in detail for use as an image generation prompt. | Instruction for the VLM. What should it say about the image? |
| system_promptopt | STRING | You are a helpful assistant that describes images concisely and accurately for use in AI image generation prompts. | System prompt for the VLM conversation |
| max_tokensopt | INT | 25632–2048 | Maximum tokens to generate |
| temperatureopt | FLOAT | 0.700–2 | Sampling temperature (0 = greedy) |
| repetition_penaltyopt | FLOAT | 1.151–2 | Penalizes repeated tokens (1.0 = off, 1.1-1.3 recommended) |
| prepend_textopt | STRING | Text to prepend before VLM output in the final result | |
| append_textopt | STRING | Text to append after VLM output in the final result |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| text | STRING | — |