Unverified50% confidenceFactExact time
COCO is a standard benchmark dataset for object detection and segmentation tasks, containing over 330,000 images and 80 categories.
1
Sources
50%
Confidence
Long-term
Relevance
10/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe project uses a self-built dataset of approximately 2,000+ annotated images, including a training set, validation set, and label annotation files68% similarUnverifiedThe dataset comes from the team's open-source NAP (Neutrae Agent Platform) project; 72 valid logs were filtered from 2,893 raw sessions, ultimately constructing 38 fine-grained scenario matrices.66% similarUnverifiedData annotation formats for object detection include COCO, YOLO, VOC, and others.65% similarUnverified最终生成的数据集包含40032个Needle格式样本，覆盖获取天气、设置定时器、开关灯三个工具，每个工具约有130种表达变体，另加40个不使用工具的负样本62% similarUnverifiedThe HybridDeepResearch benchmark contains 380 tool-dependent tasks.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/974074API
curl https://kongchang.com/api/v1/knowledge/claims/974074MCP
get_claim(id=974074)