Figure 2
From: Two-layer accumulated quantized compression for communication-efficient federated learning: TLAQC

Quantization space (b = 3).
From: Two-layer accumulated quantized compression for communication-efficient federated learning: TLAQC

Quantization space (b = 3).