takara-ai/go-attention
A pure Go implementation of attention mechanisms and transformer layers for machine learning inference.

Velocity · 7d
+1.0
★ / day
Trend
→steady
star history
Implements core transformer architecture components (attention mechanisms, transformer layers) in Go. Designed for production AI workloads with SIMD optimization, goroutine-based parallelization, and edge deployment capabilities. Targets developers wanting to run ML inference without Python or external dependencies.