MCPcopy Create free account
hub / github.com/WorldModelBench-Team/WorldModelBench / evaluate_video

Method evaluate_video

evaluation.py:139–147  ·  view source on GitHub ↗

Generate evaluation content for a video.

(self, video: 'llava.Video', prompt: str, cot: bool = True)

Source from the content-addressed store, hash-verified

137 return llava.Video(str(video_path))
138
139 def evaluate_video(self, video: 'llava.Video', prompt: str, cot: bool = True) -> str:
140 """Generate evaluation content for a video."""
141 if not cot:
142 prompt = prompt.replace(
143 "Let's think step-by-step and conclude with", "Answer with"
144 ).replace(
145 "Let's analyze step-by-step and conclude with", "Answer with"
146 )
147 return self.judge.generate_content([video, prompt])
148
149 def process_results(self, preds: Dict, accs: defaultdict) -> float:
150 """Process and print evaluation results with rich formatting."""

Callers 1

mainFunction · 0.95

Calls

no outgoing calls

Tested by

no test coverage detected