VidSTG
Unified Review
split
train
shown
10 / 10
SAM3.1 masks
模型结果 · 只读
视频切换
← 上一个视频
下一个视频 →
Token 统计 ↗
train
all lengths
shortest
middle samples
longest
any confidence
has low
has medium
all high
unlinked action participant
has actions
clear filters
train
VLM 0.5 FPS
read-only
prompt-aligned
Video evidence
keyframes
overlays
identity
模型发送给 VLM 的 clean frames;点击缩略图跳转。
帧序列
VLM · 0.5 FPS
复核 · 5 FPS
显示 grounding
清除选择
◀ Previous frame
▶ 播放
Next frame ▶
Mask identity sheet
Caption · objects & actions
Actions
action text · roles · mapped time
Entities & grounding
all mentions: click row or object card to select marks