Unverified50% confidenceBenchmarkExact time
A compact spiking neural network with 93,000 parameters has an inference energy consumption 15 times lower than the reference solution, at the cost of a 0.073 decrease in macro F1.
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe per-inference cost of hundred-billion-parameter models has dropped by an order of magnitude over the past three years.74% similarUnverified全栈优化的叠加效应使得过去18个月内每token的推理成本下降了约10倍73% similarUnverifiedFine-tuned DistilBERT achieves F1=0.9993 on in-distribution testing, but loses 12.9 mean average precision points under topic shift, a greater degradation than the linear model's 5.2 points.70% similarVerifiedINT4 quantization compresses neural network weights from FP32 to a 4-bit integer representation, reducing model size by approximately 8 times.70% similarUnverified轻量版模型单次推理的算力消耗往往只有旗舰版本的十分之一甚至更低70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/983398API
curl https://kongchang.com/api/v1/knowledge/claims/983398MCP
get_claim(id=983398)