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

<div>Normalized histogram of weights (FP32) from (a) layer 1, (b) layer 2, and (c) layer 3; transfer characteristic of the symmetric two-bits UQ for Case 1; normalized histogram of FP32 and uniformly quantized weights from (a) layer 1, (b) layer 2, and (c) layer 3.</div>

Mathematical Problems in Engineering

fig6

Figure 6

Figure 6: Performance of Post-Training Two-Bits Uniform and Layer-Wise Uniform Quantization for MNIST Dataset from the Perspective of Support Region Choice 

Figure 6 | Performance of Post-Training Two-Bits Uniform and Layer-Wise Uniform Quantization for MNIST Dataset from the Perspective of Support Region Choice 