ComfyUI Node
IF Video Prompts 🎥🧠
A ComfyUI node in ImpactFrames💥🎞️/LLM with 20 inputs and 4 outputs.
IF Video Prompts 🎥🧠
- images
- sequence_description
- scene_breakdown
- preview_image
- negative_prompt
◄input_modeFrames►
◄model_nameQwen2.5-VL-3B-Instruct-AWQ►
◄profileHyVideoAnalyzer - Simple one line prompt►
◄max_new_tokens512►
◄frame_sample_count16►
◄temperature0.70►
◄analysis_typeFull sequence►
◄languageEnglish►
◄video_file▾►
◄fps8.0►
◄max_pixels262144►
◄fallback_frame_count4►
◄custom_system_prompt►
◄prefix►
◄suffix►
◄seed-1►
◄negative_promptNone►
◄model_offloadYes►
◄precisionfloat16►
CategoryImpactFrames💥🎞️/LLM
Inputs (20)
| Name | Type | Default | Description |
|---|---|---|---|
| input_mode | COMBO | Frames | Select input mode: use pre-loaded frames or direct video file |
| model_name | COMBO | Qwen2.5-VL-3B-Instruct-AWQ | Select the Qwen2.5-VL model to use |
| profile | COMBO | HyVideoAnalyzer - Simple one line prompt | Select a profile with predefined system prompt and rules |
| max_new_tokens | INT | 5121–2048 | Maximum number of new tokens to generate |
| frame_sample_count | INT | 161–32 | Number of frames to sample from entire sequence |
| temperature | FLOAT | 0.700–2 | Higher values increase creativity but reduce coherence |
| analysis_type | COMBO | Full sequence | Type of analysis to perform on the video |
| language | COMBO | English | Language for the output |
| imagesopt | IMAGE | Input frames from a LoadVideo node | |
| video_fileopt | COMBO | Select a video file from the input directory | |
| fpsopt | FLOAT | 8.00.1–60 | Frames per second for video processing. Higher values sample more frames. |
| max_pixelsopt | INT | 2621440–921600 | Max pixels for video processing (0 = default) |
| fallback_frame_countopt | INT | 41–16 | Number of frames to use in fallback mode if initial processing fails. Lower values use less VRAM. |
| custom_system_promptopt | STRING | Custom system prompt to override the profile | |
| prefixopt | STRING | Text to add before the generated prompt | |
| suffixopt | STRING | Text to add after the generated prompt | |
| seedopt | INT | -1 | Random seed for generation (use -1 for random) |
| negative_promptopt | COMBO | None | Predefined negative prompt to use |
| model_offloadopt | COMBO | Yes | Offload model from GPU when not in use to save VRAM |
| precisionopt | COMBO | float16 | Model precision - lower precision uses less VRAM but may reduce quality |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| sequence_description | STRING | — |
| scene_breakdown | STRING | — |
| preview_image | IMAGE | — |
| negative_prompt | STRING | — |