Top-1 Accuracy (%): 85.03Weight bits: 8Activation bits: 8
Papers with Code · CC BY-SA 4.0
Quantization
Quantization is a promising technique to reduce the computation cost of neural network training, which can replace high-cost floating-point numbers (e.g., float32) with low-cost fixed-point numbers (e.g., int8/int16).
<span class="description-source">Source: Adaptive Precision Training: Quantify Back Propagation in Neural Networks with Fixed-point Numbers </span>
논문
Top-1 Accuracy (%): 78.972Weight bits: 8Activation bits: 8
Top-1 Accuracy (%): 76.4Weight bits: 8Activation bits: 8
Top-1 Accuracy (%): 77.878Weight bits: 4Activation bits: 4
Top-1 Accuracy (%): 69.79Weight bits: 4
MAP: 160327.04
Top-1 Accuracy (%): 73.8Weight bits: 4Activation bits: 4
MAP: 0.792
Top-1 Accuracy (%): 74.03
Top-1 Accuracy (%): 71.5