Unverified50% confidenceFactExact time
In the AD training stage, the wav2vec 2.0 backbone is frozen, the two task vectors are injected into each Transformer layer with per-layer learnable merging coefficients, and the entire training process optimizes only these merging coefficients and one linear classification head, with trainable parameters under 0.1% of full fine-tuning.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
Unverified一项研究提出仅训练单个Transformer层，其效果就能与全参数RL训练相媲美69% similarUnverified混合精度训练在不显著损失模型精度的前提下，可将训练速度提升约2倍，显存占用降低近一半69% similarUnverifiedLatent Compaction training is divided into two stages: the first stage fits the attention output and log normalizer of the full cache, and the second stage jointly optimizes the entire cache through end-to-end distillation.69% similarUnverifiedIn non-private training, μP has been shown to enable zero-shot learning-rate transfer from narrow to wide on Transformers.69% similarUnverifiedTransformer 架构的模型参数在训练完成后即被固定，除非进行微调或持续预训练否则无法获取新知识68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074701API
curl https://kongchang.com/api/v1/knowledge/claims/1074701MCP
get_claim(id=1074701)