The invention discloses a training method and device of an automatic driving decision model. A specific embodiment of the method comprises the following steps: in a forward calculation process of training an initial automatic driving decision model through a training sample set, determining a target arrangement matrix through a preset discrete arrangement method; selecting a target subset from a preset binary convolution kernel code set representing a complete convolution kernel space through the target arrangement matrix; determining a binary convolution kernel matrix corresponding to the initial automatic driving decision model according to the target subset and a full-precision convolution matrix of the initial automatic driving decision model; and in response to determining that the initial automatic driving decision model adopting the binary convolution kernel matrix reaches a preset condition, obtaining a trained automatic driving decision model. According to the method, the accuracy of the automatic driving decision model is ensured while the binarized automatic driving decision model is further compressed and accelerated.
本申请公开了一种自动驾驶决策模型的训练方法及装置。方法的一具体实施方式包括:在通过训练样本集训练初始自动驾驶决策模型的前向计算过程中,通过预设离散排列方法确定目标排列矩阵;通过目标排列矩阵,从表征完整卷积核空间的预设二值卷积核子码集中选择目标子集;根据目标子集和初始自动驾驶决策模型的全精度卷积矩阵,确定初始自动驾驶决策模型对应的二值卷积核矩阵;响应于确定采用二值卷积核矩阵的初始自动驾驶决策模型达到预设条件,得到训练后的自动驾驶决策模型。本申请在进一步压缩和加速二值化的自动驾驶决策模型的同时,保证了自动驾驶决策模型的准确度。
Training method and device of automatic driving decision model
自动驾驶决策模型的训练方法及装置
04.11.2022
Patent
Elektronische Ressource
Chinesisch
Training method of automatic driving decision model and vehicle control method and device
Europäisches Patentamt | 2023
|Europäisches Patentamt | 2022
|Training method of automatic driving algorithm model, automatic driving method and device
Europäisches Patentamt | 2024
|Europäisches Patentamt | 2024
|Method and device for training vehicle driving decision model
Europäisches Patentamt | 2020
|