mamba-rs 0.7.4

Mamba SSM and Mamba-3 SISO in Rust with optional CUDA acceleration: inference and training (BPTT through the SSM state, AdamW) on CPU and GPU, custom NVRTC-compiled kernels, CUDA Graph capture, f32 / bf16 / f16 storage, deterministic batch-invariant GEMMs by default with explicit cuBLAS Fast and Pedantic modes.
Documentation 
1
2
3
4
5
6

#!/bin/sh
# The same gate CI runs, before the push leaves this machine. A red gate
# here costs seconds; a red gate on the runner costs a round trip.
set -eu
cd "$(git rev-parse --show-toplevel)"
exec make ci-fast