[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fLnM0goS14ZuWBCCxHaMH0NMz5TBJ5mbfghNcEhRnGI0":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":12,"question":19,"related":20,"source":28,"type":29},[],"2025-06-06 10:08:41",186177802,[8,9,10,11],"max_pool","Conv2D","Droupout","Dense",{"count":13,"courseId":14,"courseImg":15,"courseName":16,"workId":17,"workName":18},59,"351d272eeff170c899ccc7f82b487bb2","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F5ba4e6a545b58dabb922f06d9a1383c0.jpg","机器学习基础","work_43417311","作业5 卷积神经网络","就是在不同的训练过程中随机扔掉一部分神经元.也就是让某个神经元的激活值以一定概率p让其停止工作,这次训练过程中不更新权值,也不参加神经网络的计算,但是它的权重得保留下来(只是暂时不更新而已)",[21,30,39,42,52,61,70,79,88,97],{"answer":22,"createTime":23,"id":24,"options":25,"question":27,"source":28,"type":29},[],"2025-05-29 16:08:30",186177794,[26,26,26,26],"","假设输入图像尺寸为 (32, 32, 3),使用tf.keras.layers.Conv2D(16, (3, 3), activation='relu')进行卷积操作,输出张量的形状是( )(假设默认填充方式和步长)","v1",0,{"answer":31,"createTime":5,"id":32,"options":33,"question":38,"source":28,"type":29},[],186177797,[34,35,36,37],"ModelCheckpoint","ReduceLROnPlateau","TensorBoard","CSVLogger","以下哪个回调函数可以在每个训练周期结束时保存模型的权重?( )",{"answer":40,"createTime":5,"id":6,"options":41,"question":19,"source":28,"type":29},[],[8,9,10,11],{"answer":43,"createTime":44,"id":45,"options":46,"question":51,"source":28,"type":29},[],"2025-06-05 20:04:32",186177806,[47,48,49,50],"卷积运算使用卷积核与每点周围像素的灰度值加权求和来调整该点的灰度值","卷积核一次移动的像素数称为步长","图像卷积运算的结果在计算机中以矩阵形式保存","若在图像边界不填充数字,卷积后的图像与原图像大小相同","下列关于图像卷积运算的说法,错误的是______",{"answer":53,"createTime":5,"id":54,"options":55,"question":60,"source":28,"type":29},[],186177808,[56,57,58,59],"评估模型在测试集上的性能","对训练数据进行预处理","自动调整模型的超参数","可视化训练过程和模型性能","TensorBoard 回调函数主要用于( )",{"answer":62,"createTime":5,"id":63,"options":64,"question":69,"source":28,"type":29},[],186177810,[65,66,67,68],"激活函数","卷积核大小","输入图像大小","步长","卷积神经网络隐层神经元的数量与()因素无关",{"answer":71,"createTime":5,"id":72,"options":73,"question":78,"source":28,"type":29},[],186177812,[74,75,76,77],"relu函数","随机梯度下降","以上都不对","卷积函数","下列操作中,能够在神经网络中引入非线性的是()",{"answer":80,"createTime":5,"id":81,"options":82,"question":87,"source":28,"type":29},[],186177814,[83,84,85,86],"fit()","summary()","Sequential()","compile()","回调函数是用在哪个函数中()",{"answer":89,"createTime":5,"id":90,"options":91,"question":96,"source":28,"type":29},[],186177816,[92,93,94,95],"调整学习率","记录训练日志","保存模型权重","提前终止训练,防止过拟合","在 TensorFlow 中,EarlyStopping 回调函数的主要作用是( )",{"answer":98,"createTime":99,"id":100,"options":101,"question":106,"source":28,"type":107},[],"2025-06-06 10:08:42",186177818,[102,103,104,105],"池化层","卷积层","全连接层","Dropout层","卷积神经网络的参数包括下列哪些网络层的参数",1]