Download edit_instructions/4_2.1.json from avecompass/avecompass-release: direct link, hf CLI and curl.
- Browser
- Download file 916 Bytes
-
https://huggingface.co/datasets/avecompass/avecompass-release/resolve/main/edit_instructions/4_2.1.json
- Command line
-
hf download hf://datasets/avecompass/avecompass-release/edit_instructions/4_2.1.json
-
curl -L -o 4_2.1.json https://huggingface.co/datasets/avecompass/avecompass-release/resolve/main/edit_instructions/4_2.1.json
916 Bytes
| { | |
| "video": "4_2.mp4", | |
| "task": "4_2", | |
| "instruction_index": 1, | |
| "total_instructions_for_video": 2, | |
| "instruction": { | |
| "category_label": "video_only", | |
| "category": "V2", | |
| "operation": "V2.2 Visual Crop & Zoom", | |
| "prompt_en": "Apply a 1.2x digital zoom, centering on the cat's upper body and head.", | |
| "audio_label": { | |
| "audio_op": "none", | |
| "sound_type": "n/a", | |
| "edit_aspect": "n/a" | |
| }, | |
| "difficulty": { | |
| "d1_object_localization": "hard", | |
| "d1_reason": "The instruction requires locating and centering on a specific part of the cat (upper body and head) within the frame, which is not a global operation nor an extreme close-up filling most of the frame.", | |
| "d2_audio_complexity": "simple", | |
| "d3_cross_modal_linkage": "none", | |
| "d3_reason": "This is a pure video-only edit with no audio operation, so no cross-modal inference is required." | |
| } | |
| } | |
| } | |