GQLA: Group-Query Latent Attention for Hardware-Adaptive Decoding

Name: GQLA: Group-Query Latent Attention for Hardware-Adaptive Decoding
Uploaded: 2026-10-17T17:45:45.162280+00:00
Duration: 350 s
Description: Group-Query Latent Attention (GQLA) introduces two mathematically equivalent decoding paths—an MQA-absorb path and a Group-Query Attention path—using a single set of trained weights. This allows runtime selection of the hardware-matched path without retraining, enabling models to adapt efficiently to different GPU architectures like H100 or H20.

Group-Query Latent Attention (GQLA) introduces two mathematically equivalent decoding paths—an MQA-absorb path and a Group-Query Attention path—using a single set of trained weights. This allows runtime selection of the hardware-matched path without retraining, enabling models to adapt efficiently to different GPU architectures like H100 or H20.

Open the full episode