Unverified50% confidenceFactExact time
After LoRA training is complete, BA can be merged back into W (W' = W + BA), so there is no additional latency at inference time and it can be completely consistent with the original model structure.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDuring LoRA inference, W' = W₀ + (α/r)BA can be pre-computed to merge the LoRA branch into the original weights, achieving zero additional inference latency.75% similarUnverifiedDuring LoRA training, the pre-trained model weights W₀ are frozen and only the two matrices B and A are trained, with the forward propagation formula H = W₀X + (α/r)BAX.74% similarUnverifiedLoRA推理时可将BA与原权重合并，不增加任何推理延迟，通常只应用于注意力层的Q、V权重矩阵69% similarVerifiedWith LoRA, the weights can be merged back into the original model at inference time, introducing no additional latency68% similarUnverifiedLoRA 由 Hu 等人在 2021 年提出，通过为每层注入可训练的低秩分解矩阵实现参数更新 ΔW = BA，将可训练参数量压缩至原始模型的 0.1%-1%66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1047749API
curl https://kongchang.com/api/v1/knowledge/claims/1047749MCP
get_claim(id=1047749)