Unverified50% confidenceFactExact time
V1–V4无法恢复已被排除，V6是早期不稳定的基线无法精确聚合，因此99.3%是在可重建、可评分的子集上得出的
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/24/2026
First Seen
Valid until: 11/22/2026
Sources
Related Entities
Related Claims
UnverifiedIn the Q8_K_XL scheme, MoE layers retain the original MXFP4, while the remaining layers that were originally FP8 are upcast to BF16, and this conversion is 100% lossless60% similarUnverifiedCgroups v2 was introduced in Linux 4.5, adopting a unified hierarchy structure to replace the multi-hierarchy design of v1.56% similarUnverifiedV100 的 Volta 架构 Tensor Core 仅支持 FP16/FP32 混合精度，不具备原生 FP8 矩阵乘法指令，FP8在V100上无法获得算力加速56% similarUnverified在非50系显卡上使用NVFP4会退化为软件模拟，导致显存占用大、精度差，无任何优势55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/795251API
curl https://kongchang.com/api/v1/knowledge/claims/795251MCP
get_claim(id=795251)