ComfyUI Node Runs on cloud
MiniCPM-V Model
A ComfyUI node in VLM Nodes/MiniCPM-V with 8 inputs and 1 output.
MiniCPM-V Model
- image
- STRING
◄promptDescribe this image in detail.►
◄model_variant▾►
◄context_length4096►
◄temperature0.7►
◄top_p0.8►
◄top_k100►
◄repeat_penalty1.05►
CategoryVLM Nodes/MiniCPM-V
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | Input image to be analyzed by MiniCPM-V | |
| prompt | STRING | Describe this image in detail. | Instructions for the model. Be specific about what aspects of the image you want analyzed. |
| model_variant | COMBO | Model size/quality tradeoff. Smaller models (Q2-Q4) are faster but less accurate. Larger models (Q8, F16) provide better quality but require more VRAM. | |
| context_length | INT | 4096512–8192 | Maximum length of text context. Larger values allow longer conversations but use more memory. Default 4096 works well for most cases. |
| temperature | FLOAT | 0.70.1–2 | Controls randomness in generation. Lower values (0.1-0.5) are more focused and deterministic. Higher values (0.8-2.0) increase creativity and variance. |
| top_p | FLOAT | 0.80.1–1 | Nucleus sampling threshold. Lower values make responses more focused. Higher values allow more diverse word choices. |
| top_k | INT | 1001–1000 | Limits the number of tokens considered for each generation step. Lower values increase focus, higher values allow more variety. |
| repeat_penalty | FLOAT | 1.051–2 | Penalizes word repetition. Values above 1.0 discourage repeated phrases. Higher values (>1.3) may affect fluency. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| STRING | STRING | — |