ComfyUI Node
FL VLM Image Inference
A ComfyUI node in FL/VLM with 19 inputs and 2 outputs.
FL VLM Image Inference
- vlm_model
- image
- response
- metadata_json
◄system_promptYou are a precise vision-language assistant. Answer only from the provided image content.►
◄promptDescribe the image in detail.►
◄batch_modeall_images_one_prompt►
◄image_start0►
◄image_stride1►
◄max_images16►
◄min_pixels64►
◄max_pixels256►
◄max_new_tokens512►
◄do_samplefalse►
◄temperature0.70►
◄top_p0.90►
◄top_k50►
◄repetition_penalty1.00►
◄seed-1►
◄cleanup_temp_imagestrue►
◄resize_preset360p_balanced►
CategoryFL/VLM
Inputs (19)
| Name | Type | Default | Description |
|---|---|---|---|
| vlm_model | FL_VLM_MODEL | — | |
| image | IMAGE | — | |
| system_prompt | STRING | You are a precise vision-language assistant. Answer only from the provided image content. | — |
| prompt | STRING | Describe the image in detail. | — |
| batch_mode | COMBO | all_images_one_prompt | 2 options: all_images_one_prompt, each_image_separately |
| image_start | INT | 00–100000 | — |
| image_stride | INT | 11–100000 | — |
| max_images | INT | 161–512 | — |
| min_pixels | INT | 6464–4096 | Qwen visual token budget lower bound. Internally multiplied by 28*28. |
| max_pixels | INT | 25664–8192 | Qwen visual token budget upper bound for manual_qwen_budget. Internally multiplied by 28*28. |
| max_new_tokens | INT | 5121–8192 | — |
| do_sample | BOOLEAN | false | — |
| temperature | FLOAT | 0.700–5 | — |
| top_p | FLOAT | 0.900–1 | — |
| top_k | INT | 500–500 | — |
| repetition_penalty | FLOAT | 1.000.1–5 | — |
| seed | INT | -1-1–18446744073709550000 | — |
| cleanup_temp_images | BOOLEAN | true | — |
| resize_preset | COMBO | 360p_balanced | Controls Qwen's image resize budget before inference. Existing min_pixels/max_pixels are only used by manual_qwen_budget. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| response | STRING | — |
| metadata_json | STRING | — |