http://blog.caveduck.io/tags/int8/Recent content in INT8 on Warpspace BlogHugo -- 0.154.5en-USMon, 20 Apr 2026 00:00:00 +0000http://blog.caveduck.io/posts/neural-net-quantization/Mon, 20 Apr 2026 00:00:00 +0000http://blog.caveduck.io/posts/neural-net-quantization/We take the trained 32-bit weights of a neural net and shrink them down to 2–8 bits using two approaches: K-Means (non-uniform) and Linear (uniform, integer arithmetic). From the affine mapping r=S(q−Z) all the way to the compression-ratio vs. accuracy trade-off, with plots measured by actually running the code on an MNIST MLP.