Instruction-Tuned From survey PDF — verify links
Holmes-VAD
Holmes-VAD — Anomaly Understanding
arXiv · 2024Human-reviewed summary status
Survey summary
Holmes-VAD builds VAD-Instruct50k, a large-scale multimodal instruction-tuning benchmark created via semi-automatic labeling with a video captioner and an LLM, then trains a lightweight temporal sampler to select high-anomaly-response frames and fine-tunes an MLLM for anomaly localization and explanation.
Taxonomy placement
Instruction-Tuned
Instruction-TunedAnomaly Understanding
Method characteristics
Instruction-TunedVADInstruct50KTemporal samplerLoRA