Performance of Post-Training Two-Bits Uniform and Layer-Wise Uniform Quantization for MNIST Dataset from the Perspective of Support Region Choice

Table 3

Accuracy comparison of compressed NN models.

	1-bit [15]	2-bits UQ [20]	2-bits UQ [21]	2-bits adaptive UQ [31]	Our Case 1	Our layer-wise Case 1
Accuracy (%)	91.12	94.70	94.49	96.26	96.97	97.26