Unverified50% confidenceBenchmarkExact time
CAPI-DINO的ViT-B/14模型在iNat21数据集上取得了68.7%的线性探测准确率
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/9/2026
First Seen
Valid until: 12/8/2026
Sources
Related Entities
Related Claims
Unverified该量化模型相较于BF16全精度版本保留了约77%的准确率73% similarUnverified在同一细粒度汽车分类任务中，CLIP ViT-L的k-NN准确率约为59%65% similarUnverifiedPathway报告称基于BDH的系统在约25万道Sudoku Extreme题目上达到97.4%的Top-1准确率，且未使用CoT、未进行解答回溯65% similarUnverified在细粒度汽车分类的k-NN实验中，SigLIP2 SO400M达到约92%准确率，CLIP ViT-L约59%，DINOv2 Giant约41%64% similarUnverified微调后的Qwen3.5-4B模型在BU Bench V1基准上准确率从15%提升至53%62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/884098API
curl https://kongchang.com/api/v1/knowledge/claims/884098MCP
get_claim(id=884098)