Unverified50% confidenceFactExact time
AWS Aurora通过6副本4/6写入法定人数（跨3个可用区各2副本）实现持久性，声称年化数据丢失率低于10^-12
1
Sources
50%
Confidence
Long-term
Relevance
8/5/2026
First Seen
Sources
Related Claims
UnverifiedINT8 quantization loses approximately 0.5% accuracy while halving VRAM usage; INT4 quantization compresses VRAM to 1/4 of the original with an accuracy loss of approximately 1–3%.61% similarUnverified将年龄字段从100个可能值泛化为5个区间，信息熵从约6.64比特降至约2.32比特，信息量损失超过60%59% similarUnverified4-bit quantization reduces model size to 1/8 of the original at the cost of approximately 1–3% accuracy loss.56% similarUnverified研究表明将收入从精确值泛化为5个区间可能导致模型AUC下降2-5个百分点54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/691095API
curl https://kongchang.com/api/v1/knowledge/claims/691095MCP
get_claim(id=691095)