[KongchangAI]
ConceptFlashTopK算子

FlashTopK

An efficient TopK selection operator for sparse attention mechanisms. It substantially reduces the time consumed by the attention computation bottleneck and improves the inference efficiency of large models.

Timeline (last 90 days)

Nov 14

Tencent Hunyuan, together with the University of Edinburgh, officially open-sourced the FlashTopK operator, aimed at accelerating the TopK filtering bottleneck in DeepSeek-style sparse attention.

Unverified50%
Nov 14

FlashTopK reduces the time spent on the TopK filtering step from 5.32 seconds in the traditional implementation to 1.62 seconds, a reduction of about 70%.

Unverified50%
Nov 14

According to measured data from the repository, FlashTopK is 1.7 to 3.8 times faster than comparable mainstream implementations in the inference stage, and 1.3 to 2 times faster in absorption training.

Unverified50%
Nov 14

FlashTopK is currently adapted to the GLM5 series as well as the Hunyuan Hi4 preview version.

Unverified50%

All Facts (4)

Source Articles