Unverified50% confidenceFactExact time
SmoothQuant (2022) smoothly transfers outliers in activations to weights through a mathematically equivalent transformation, enabling both activations and weights to be quantized using INT8.
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
Unverified旋转变换量化的理论基础是正交矩阵不改变模型数学等价性但能重新分布数值统计特性，用于搅匀离群值63% similarUnverifiedThe parameters of logistic regression are estimated by maximizing the log-likelihood function (equivalent to minimizing the cross-entropy loss).63% similarUnverifiedDiffusionGemma是首个在大规模（26B参数）上证明扩散范式可以匹敌自回归质量的工作61% similarUnverifiedFull fine-tuning updates all weights of the model, yielding the best results but requiring extremely high computing power61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/829121API
curl https://kongchang.com/api/v1/knowledge/claims/829121MCP
get_claim(id=829121)