Video Question Answering

Best model: VideoChat2

Best model: GF (sup) - Faster RCNN

Best model: Text + Text (no Multimodal Pretext Training)

Best model: TimeSformer

Best model: VideoChat2_mistral

Best model: FrozenBiLM

Best model: Clover

Best model: VIOLETv2

Best model: ATP (1<-16)

Best model: Singularity-temporal

Best model: FrozenBiLM

Best model: Tarsier (34B)

Best model: LinVT-Qwen2-VL (7B)

Best model: ViLA (3B, 4 frames)

Best model: Oyrx (34B)

Best model: GIT

Best model: VLAP (4 frames)

Best model: Tarsier-34B

Best model: LLaMA-VQA

Best model: Just Ask (fine-tune)