Hugging Face Blog·· 2024-08-14AI 评分47
Hugging Face 复现 Infini-Attention 失败:为何仍应继续尝试?
A failed experiment: Infini-Attention, and why we should keep trying?
AI 导读
Hugging Face 复现 Google 的 Infini-attention 论文后发现,压缩记忆次数越多,性能反而越差,90% 的时间花在调试收敛问题上。
来源:Hugging Face Blog · huggingface.co