Neural network generation and image processing method and device, platform and electronic equipment

Abstract

The embodiment of the invention discloses a neural network generation and image processing method and device, a platform and electronic equipment, and the method comprises the steps: obtaining the fixed-point operation resource information of a hardware platform which is expected to be deployed by a neural network; adjusting the distribution of the values of the network parameters of at least onenetwork unit included in the neural network to be uniform distribution; training the neural network after parameter adjustment based on the sample image data; subjecting the values of the network parameters in the trained neural network to fixed-point adjustment according to the fixed-point operation resource information, reducing the values of the network parameters adjusted to be uniformly distributed are more suitable for fixed-point adjustment, the precision loss caused by conversion from the floating-point number to the fixed-point number, and improving the precision of the neural networksubjected to fixed-point adjustment; and carrying out fixed-point adjustment out based on the fixed-point operation resource information of the hardware platform, so that the adjusted neural networkcan run on the platform with limited hardware resources.

Description

technical field [0001] The present disclosure relates to computer vision technology, in particular to a neural network generation and image processing method and apparatus, platform, and electronic device. Background technique [0002] With the development of deep convolutional neural networks, deeper or wider network structures continue to refresh the accuracy of various computer vision datasets, however, the deepening or widening of the network is bound to increase the running time of the network in the prediction process. In recent years, in order to deploy deep convolutional neural networks on resource platforms such as FPGAs with low power consumption and suitable for fixed-point computing, more and more researchers have begun to study the quantization of floating-point models trained on GPUs and other devices. into a low-bit fixed-point model. The designed fixed-point model makes the model smaller and still guarantees the accuracy, which has become a hot direction in ...

AI Technical Summary

Problems solved by technology

[0002] With the development of deep convolutional neural networks, deeper or wider network structures continue to refresh the accuracy of various com...

Images