Neural network searching method and system
A neural network and search method technology, applied in neural learning methods, biological neural network models, etc., can solve problems such as inability to obtain search results, poor generalization ability of neural network search methods, etc., to alleviate unfair training and achieve good classification results , efficient and robust training
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0033] The embodiment of the present invention provides a neural network search method, such as figure 1 As shown, the method includes the following steps:
[0034] Step S1: Randomly sample two non-overlapping models on the super-network to be searched, the two sampled models have one copy each in the encoder and the momentum encoder, and train the encoder based on the momentum contrastive learning method.
[0035] In the embodiment of the present invention, the initial structure of the momentum encoder is the same as that of the corresponding encoder. During the training process, the update of the parameters of the momentum encoder is only based on the parameters of the encoder, and the encoder is trained unsupervised by momentum comparison learning. The specific process of training the encoder based on the momentum contrastive learning method, such as figure 2 As shown, the momentum encoder is initialized using the same network structure as the encoder, and in each trainin...
Embodiment 2
[0044] The embodiment of the present invention provides a neural network search system, such as image 3 shown, including:
[0045] Sampling model training module 1 is used to randomly sample two non-overlapping models on the super-network to be searched. The two sampled models have one copy each in the encoder and the momentum encoder. The method based on momentum contrastive learning compares the encoder and the momentum encoder. The momentum encoder is trained; this module executes the method described in step S1 in Embodiment 1, and details are not repeated here.
[0046] Classification head training module 2, used to take out the trained encoder and fix its network weights, set all batch normalization layers in the encoder to training mode, and add a randomly initialized fully connected classification head after the encoder structure part, and train the classification head; this module executes the method described in step S2 in Embodiment 1, and details are not repeated...
Embodiment 3
[0051] An embodiment of the present invention provides a computer device, such as Figure 4 As shown, the device may include a processor 51 and a memory 52, wherein the processor 51 and the memory 52 may be connected by a bus or in other ways, Figure 4 Take connection via bus as an example.
[0052] The processor 51 may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-available processors A programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components and other chips, or a combination of the above-mentioned types of chips.
[0053] As a non-transitory computer-readable storage medium, the memory 52 can be used to store non-transitory software programs, n...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More