tensorflow

    1热度

    1回答

    我正在使用docker来运行tensorflow和再培训启动模块。我使用下面的代码: docker run -it \ --publish 6006:6006 \ --volume ${HOME}/tf_files:/tf_files \ --workdir /tf_files \ tensorflow/tensorflow:1.1.0 bash 然后我用 python retrain

    0热度

    1回答

    我遵循专家的tensorflow MNIST教程。我写下如下所示的代码,这是本教程的副本。但是,当我运行我的代码时,其准确率仅为92%,86%......它在我的Mac上运行速度仅为1或2分钟。而随着步骤的增加,精度 step 0, training accuracy 0.08 step 100, training accuracy 0.1 step 200, training accurac

    0热度

    2回答

    我想创建一个只有一个类的tensorflow中的简单逻辑回归模型。但是,由于某些原因,tf.sigmoid函数返回的是数组类型而不是标量。 的类型从成本函数返回值是一个np.ndarray() 形状(3390,2) 我想不通为什么tf.sigmoid函数将返回数组类型而不是标量... 任何帮助表示赞赏。 这里是我的代码: #!/usr/bin/env python3 import tensor

    1热度

    1回答

    我想用TensorFlow作为后端运行Keras,但我想在GPU上运行它。我安装TensorFlow-GPU,CUDA 8.0,cuDNN,但Keras得到以下错误: --------------------------------------------------------------------------- ImportError Traceback (most rece

    0热度

    1回答

    TensorFlow 1.4文档提供了演示Dataset.from_generator的用法的代码。当我运行的代码,我得到一个InvalidArgumentError:0-th value returned by pyfunc_0 is int32, but expects int64。 我使用Python 3.6.1。这里的代码: def gen(): for i in itertoo

    0热度

    1回答

    def model_fn(features, labels, mode, params): """Model function for Estimator.""" # Connect the first hidden layer to input layer # (features["x"]) with relu activation first_hidd

    1热度

    1回答

    Tensorflow中默认的全球variable_scope是什么?我如何检查对象?有没有人有关于此的想法?

    0热度

    1回答

    我有一个由图像组成的训练集,从女巫我必须预测由固定数量的字母和数字组成的标签。 将这些标签添加到张量流中的最佳方式是什么?我想创建一个数字列表,其中一个用于标签中的所有字符/数字。 我创建了一个清单,阿拉的可能值: __dict = ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q

    0热度

    1回答

    我工作在tensorflow实施新TFGAN模块TFGAN Module 有没有人能够真正得到它的工作?我运行到经过tf.random.noise成一个简单的发电机的问题: tfgan = tf.contrib.gan noise = tf.random_normal([BATCH_SIZE, 28,28]) def my_generator(z, out_dim=28*28, n_unit

    1热度

    2回答

    我正在使用TensorFlow处理与CNN相关的项目。 我输入使用图像(20个这样的图像) for filename in glob.glob('input_data/*.jpg'): input_images.append(cv2.imread(filename,0)) image_size_input = len(input_images[0]) 这些图像是尺寸(250250),因为