摘要：【中文】本申请公开了障碍物识别方法、装置、电子设备及存储介质，涉及自动驾驶领域，其中方法可包括：在主车行驶过程中，获取路侧感知信息，所述主车为待识别的车辆；针对获取到的任一帧路侧感知信息，分别进行以下处理：若当前未处于所定义的异常状态，则根据当前帧对应的主车信息识别出路侧感知信息中的主车障碍物，并缓存当前帧对应的主车信息；若当前处于异常状态，则根据最新缓存的主车信息估算出当前帧对应的主车信息，根据当前帧对应的主车信息识别出路侧感知信息中的主车障碍物，并缓存当前帧对应的主车信息。应用本申请所述方案，可提升主车障碍物识别的鲁棒性和容错性等。 【EN】The application discloses an obstacle identification method, an obstacle identification device, electronic equipment and a storage medium, and relates to the field of automatic driving, wherein the method comprises the following steps: the method comprises the steps that roadside perception information is obtained in the driving process of a main vehicle, wherein the main vehicle is a vehicle to be identified; aiming at any one frame of acquired roadside perception information, the following processing is respectively carried out: if the current frame is not in the defined abnormal state, recognizing a main vehicle barrier in the roadside perception information according to the main vehicle information corresponding to the current frame, and caching the main vehicle information corresponding to the current frame; and if the current state is abnormal, estimating the host vehicle information corresponding to the current frame according to the host vehicle information cached latest, identifying the host vehicle barrier in the roadside perception information according to the host vehicle information corresponding to the current frame, and caching the host vehicle information corresponding to the current frame. By applying the scheme, the robustness, the fault tolerance and the like of the main vehicle obstacle identification can be improved.

摘要：【中文】本公开涉及人工智能领域。本公开的实施例公开了模型生成方法和装置。该方法包括：通过依次执行多次迭代操作生成用于执行深度学习任务的神经网络模型；其中，迭代操作包括：基于当前的奖励反馈值更新神经网络模型的结构，其中，奖励反馈值的初始值是预设的数值；根据神经网络模型的当前量化策略，对更新后的神经网络模型进行训练；获取训练后的神经网络模型的性能，并根据训练后的神经网络模型的性能更新奖励反馈值；响应于确定奖励反馈值达到预设的收敛条件或迭代操作的次数达到预设的阈值，确定当前迭代操作中训练后的神经网络模型为用于执行深度学习任务的神经网络模型。该方法提升了神经网络模型的运算效率。 【EN】The present disclosure relates to the field of artificial intelligence. The embodiment of the disclosure discloses a model generation method and a model generation device. The method comprises the following steps: generating a neural network model for performing a deep learning task by sequentially performing a plurality of iterative operations; wherein the iterative operation comprises: updating the structure of the neural network model based on the current reward feedback value, wherein the initial value of the reward feedback value is a preset numerical value; training the updated neural network model according to the current quantization strategy of the neural network model; acquiring the performance of the trained neural network model, and updating the reward feedback value according to the performance of the trained neural network model; and in response to the fact that the reward feedback value reaches a preset convergence condition or the number of times of the iterative operation reaches a preset threshold value, determining the neural network model trained in the current iterative operation as the neural network model for executing the deep learning task. The method improves the operation efficiency of the neural network model.

摘要：【中文】本公开涉及人工智能领域。本公开的实施例公开了模型生成方法和装置。该方法包括：通过依次执行多次迭代操作生成用于执行深度学习任务的神经网络模型；迭代操作包括：基于当前的奖励反馈值，在与预设的神经网络模型的各网络结构单元分别对应的量化方法搜索空间中确定出各网络结构单元的当前量化方法，以更新预设的神经网络模型的量化策略；基于更新后的量化策略对预设的神经网络模型进行量化；获取量化后的神经网络模型的性能，并更新奖励反馈值；响应于确定奖励反馈值达到预设的收敛条件或迭代操作的次数达到预设的阈值，确定当前量化后的神经网络模型为生成的用于执行深度学习任务的神经网络模型。该方法可以减少神经网络模型占用的内存空间。 【EN】The present disclosure relates to the field of artificial intelligence. The embodiment of the disclosure discloses a model generation method and a model generation device. The method comprises the following steps: generating a neural network model for performing a deep learning task by sequentially performing a plurality of iterative operations; the iterative operation comprises: determining the current quantization method of each network structure unit in a quantization method search space respectively corresponding to each network structure unit of the preset neural network model based on the current reward feedback value so as to update the quantization strategy of the preset neural network model; quantizing the preset neural network model based on the updated quantization strategy; acquiring the performance of the quantized neural network model, and updating the reward feedback value; and in response to determining that the reward feedback value reaches a preset convergence condition or the number of iterative operations reaches a preset threshold, determining the currently quantized neural network model as the generated neural network model for executing the deep learning task. The method can reduce the memory space occupied by the neural network model.

摘要：【中文】本申请公开了一种信息推荐方法、装置以及电子设备，涉及信息推荐领域。具体实现方案为：根据用户的与经济状态有关的多种属性信息，及各属性信息对应的属性参数；根据各属性信息对应的属性参数计算用户的经济状态参数；比较经济状态参数和经济状态阈值，根据比较结果确定用户的经济状态标签；根据用户的经济状态标签向用户推荐信息。根据用户的经济状态标签可以更准确更全面的给用户推荐信息。 【EN】The application discloses an information recommendation method and device and electronic equipment, and relates to the field of information recommendation. The specific implementation scheme is as follows: according to various attribute information of the user related to the economic state and attribute parameters corresponding to the attribute information; calculating economic state parameters of the user according to the attribute parameters corresponding to the attribute information; comparing the economic state parameter with an economic state threshold value, and determining an economic state label of the user according to a comparison result; and recommending information to the user according to the economic state label of the user. The information can be more accurately and comprehensively recommended to the user according to the economic state label of the user.

摘要：【中文】本申请公开了一种训练样本生成方法、装置以及电子设备，涉及训练样本领域。具体实现方案为：获取初始样本编码序列及其对应的多个均衡属性的标签；将初始样本编码序列、编码序列长度、多个均衡属性的标签输入至自适应均衡生成模型中，得到预测样本编码序列；根据预测样本编码序列计算均衡值，并利用均衡值对自适应均衡生成模型进行更新；在自适应均衡生成模型收敛的情况下，得到均衡样本编码序列，对均衡样本编码序列进行解码，得到均衡样本集合。利用均衡样本集合训练深度神经元网络模型，达到了不仅加快模型的收敛速度，而且提升模型精度的技术效果。 【EN】The application discloses a training sample generation method and device and electronic equipment, and relates to the field of training samples. The specific implementation scheme is as follows: acquiring an initial sample coding sequence and a plurality of corresponding balanced attribute labels thereof; inputting the initial sample coding sequence, the coding sequence length and a plurality of labels with balanced attributes into a self-adaptive balanced generation model to obtain a predicted sample coding sequence; calculating an equilibrium value according to the prediction sample coding sequence, and updating the self-adaptive equilibrium generation model by using the equilibrium value; and under the condition that the self-adaptive equalization generation model is converged, obtaining an equalization sample coding sequence, and decoding the equalization sample coding sequence to obtain an equalization sample set. The deep neural network model is trained by utilizing the balanced sample set, so that the technical effects of not only accelerating the convergence speed of the model, but also improving the precision of the model are achieved.

摘要：【中文】本申请公开了一种候选框过滤方法、装置以及电子设备，涉及候选框过滤领域。具体实现方案为：将超参数的个数和超参数范围输入至过滤策略生成模型，得到超参数序列；根据超参数序列，从待检测图片中的全部候选框中过滤掉冗余的候选框；根据保留的候选框计算超参数序列对应的检测评价信息；利用检测评价信息更新过滤策略生成模型，直至收敛，得到目标候选框。解决基于人工设置非极大抑制阈值方法不能同时保证检测框准确率和召回率的问题，达到同时提高目标检测任务的准确率和召回率的技术效果。 【EN】The application discloses a candidate frame filtering method and device and electronic equipment, and relates to the field of candidate frame filtering. The specific implementation scheme is as follows: inputting the number of the hyper-parameters and the hyper-parameter range into a filtering strategy generation model to obtain a hyper-parameter sequence; filtering redundant candidate frames from all candidate frames in the picture to be detected according to the hyper-parameter sequence; calculating detection evaluation information corresponding to the hyper-parameter sequence according to the reserved candidate frame; and updating the filtering strategy generation model by using the detection evaluation information until convergence to obtain the target candidate frame. The method solves the problem that the accuracy and the recall rate of the detection frame cannot be ensured simultaneously based on a method for manually setting the non-maximum inhibition threshold, and achieves the technical effect of simultaneously improving the accuracy and the recall rate of the target detection task.

摘要：【中文】本申请公开了一种拍摄盲区的长度确定方法、装置以及计算机设备，涉及智能交通技术领域。具体实现方案为：通过获取第一摄像头和第二摄像头同步采集的初始图像帧；根据第一摄像头采集的初始图像帧，确定可视区域的边界；从第二摄像头采集的初始图像帧中，确定目标车辆；根据第一摄像头后续采集的图像帧，监测目标车辆行驶至边界位置之前，边界位置通过的各行驶车辆；根据各行驶车辆，确定第一摄像头和第二摄像头的拍摄盲区长度。该方法通过目标车辆行驶至可视区域的边界位置之前，边界位置通过的各行驶车辆确定第一摄像头和第二摄像头的拍摄盲区长度，避免了相关技术中人工测量拍摄盲区长度，导致测量成本高、测量准确度较低的技术问题。 【EN】The application discloses a method and a device for determining the length of a shooting blind area and computer equipment, and relates to the technical field of intelligent traffic. The specific implementation scheme is as follows: acquiring initial image frames synchronously acquired by a first camera and a second camera; determining the boundary of a visual area according to an initial image frame acquired by a first camera; determining a target vehicle from the initial image frame collected by the second camera; monitoring running vehicles passing through the boundary position before the target vehicle runs to the boundary position according to image frames subsequently acquired by the first camera; and determining the lengths of the shooting blind areas of the first camera and the second camera according to the running vehicles. According to the method, the lengths of the shooting blind areas of the first camera and the second camera are determined by the running vehicles passing the boundary positions before the target vehicle runs to the boundary position of the visible area, so that the technical problems of high measurement cost and low measurement accuracy caused by manual measurement of the lengths of the shooting blind areas in the related technology are solved.

摘要：【中文】根据本公开的示例实施例，提供了图像处理方法、装置、设备和计算机可读存储介质，可用于人工智能领域。图像处理方法包括基于样本图像集中的样本图像所包括的通道，确定针对样本图像集的目标通道集合，目标通道集合中的通道具有不同的类型。方法还包括基于目标通道集合，确定用于图像处理模型的通道参数，通道参数指示与图像处理模型对样本图像所执行操作相对应的通道数目。方法进一步包括基于通道参数和样本图像集，生成图像处理模型。所得到的图像处理模型能够处理具有任意数目通道的图像。以此方式，能够利用不同通道所提供的信息，从而提高图像处理结果的准确性。 【EN】According to an example embodiment of the present disclosure, an image processing method, an apparatus, a device, and a computer-readable storage medium are provided, which may be used in the field of artificial intelligence. The image processing method comprises the step of determining a target channel set aiming at a sample image set based on channels included in sample images in the sample image set, wherein the channels in the target channel set are of different types. The method also includes determining, based on the target set of channels, a channel parameter for the image processing model, the channel parameter indicating a number of channels corresponding to an operation performed by the image processing model on the sample image. The method further includes generating an image processing model based on the channel parameters and the sample image set. The resulting image processing model is capable of processing images having any number of channels. In this way, the information provided by different channels can be utilized, thereby improving the accuracy of the image processing results.

摘要：【中文】本申请实施例公开了一种显存处理方法、装置、设备和介质，涉及计算机领域，尤其涉及显存的管理技术。具体实现方案为：基于显存变量的显存空间，执行所述显存变量的运算逻辑；若检测到所述显存变量的生命周期结束，则调用所述显存变量的显存释放析构函数，释放所述显存变量的显存空间。本申请实施例提供了一种显存处理方法、装置、设备和介质，实现了对显存的自动释放，帮助开发人员避免手动释放显存的麻烦。 【EN】The embodiment of the application discloses a method, a device, equipment and a medium for processing a video memory, relates to the field of computers, and particularly relates to a video memory management technology. The specific implementation scheme is as follows: executing the operation logic of the video memory variable based on the video memory space of the video memory variable; and if the end of the life cycle of the video memory variable is detected, calling a video memory release destructor of the video memory variable, and releasing the video memory space of the video memory variable. The embodiment of the application provides a method, a device, equipment and a medium for processing a video memory, which realize automatic release of the video memory and help developers to avoid the trouble of manually releasing the video memory.

摘要：【中文】本申请公开了模型迁移方法和电子设备，涉及机器学习技术领域。具体实现方案为：对第一平台上的学习模型的第一数据文件进行解析，获得所述学习模型的网络结构；所述网络结构包括N个节点，N为正整数；根据所述N个节点的拓扑序列，依次将所述N个节点映射到第二平台，获得M个节点，M为正整数，M大于或等于N；根据所述M个节点，生成所述学习模型在所述第二平台上的第二数据文件。可将第一平台的学习模型迁移至第二平台上，无需重新编写代码，也不用对迁移至第二平台上的学习模型进行重新训练，节省重新编码和重新训练的时间，提高了学习模型的迁移效率。 【EN】The application discloses a model migration method and electronic equipment, and relates to the technical field of machine learning. The specific implementation scheme is as follows: analyzing a first data file of a learning model on a first platform to obtain a network structure of the learning model; the network structure comprises N nodes, wherein N is a positive integer; according to the topological sequence of the N nodes, sequentially mapping the N nodes to a second platform to obtain M nodes, wherein M is a positive integer and is greater than or equal to N; and generating a second data file of the learning model on the second platform according to the M nodes. The learning model of the first platform can be transferred to the second platform without rewriting codes or retraining the learning model transferred to the second platform, so that the time for recoding and retraining is saved, and the transfer efficiency of the learning model is improved.