ComfyUI Node
MiniCPM-V Q&A
A ComfyUI node in MiniCPM-V with 9 inputs and 1 output.
MiniCPM-V Q&A
- minicpmv_model
- image
- STRING
◄questionDescribe this scene in great detail in English►
◄path►
◄n_frames4►
◄max_tokens512►
◄seed0►
◄temperature1.0►
◄force_offloadtrue►
CategoryMiniCPM-V
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| minicpmv_model | MINICPMV_MODEL | — | |
| question | STRING | Describe this scene in great detail in English | Question or prompt for the model to analyze the image(s) |
| imageopt | IMAGE | Input image or batch of frames | |
| pathopt | STRING | Path to: video file, image file, or folder of images | |
| n_framesopt | INT | 40–64 | Frames to sample. 0: all frames (may OOM), 1: middle, 2: first/last, 3: first/mid/last, 4+: evenly spaced |
| max_tokensopt | INT | 5121–2048 | Maximum length of generated response |
| seedopt | INT | 00–18446744073709550000 | Seed for random generation (0 for random) |
| temperatureopt | FLOAT | 1.00–2 | Higher values make output more random, lower values more deterministic |
| force_offloadopt | BOOLEAN | true | If true, the model will be offloaded to save memory after generation |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| STRING | STRING | — |