Unverified50% confidenceBenchmarkExact time
IQ4_NL 相比 FP16 可节省约 75% 显存
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
UnverifiedO4-mini achieved an accuracy rate exceeding 80% on the GPQA doctoral-level question-answering benchmark.62% similarUnverifiedFor normally distributed data, the 1.5×IQR range covers approximately 99.3% of data points.60% similarUnverified通过INT8或INT4量化，可将模型显存占用降低50%至75%60% similarUnverifiedOn the CLINC150 benchmark, 80 fine-tuning samples can bring vector routing to an accuracy of 85.9%, while a full fine-tune can exceed 96%.59% similarUnverifiedFP8相比BF16可实现近2倍的吞吐量提升，同时显存占用减少约50%58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/537825API
curl https://kongchang.com/api/v1/knowledge/claims/537825MCP
get_claim(id=537825)