Technologies

Echo

An error-compensated 4-bit all-reduce method that speeds up AI model training and inference without special hardware.

1 story, first covered on Oct 14, 2026.

Every story about Echo (1)

A new communication method cuts the data-hauling bottleneck in large AI training by sending just 4 bits at a time, without breaking model quality.

Oct 14, 2026