Question

我正在尝试使用Tensorflow 1.1.0和TFLearn 0.3.1为大量癌症图像（.png）构建自定义CNN分类器，主要是跟随其他人的CNN分类器here，但是当我尝试为了适应我的模型，Tensorflow会抛出以下错误：

W tensorflow/core/framework/op_kernel.cc:983] Unimplemented: Cast string to float is not supported
E tensorflow/core/common_runtime/executor.cc:594] Executor failed to create kernel. Unimplemented: Cast string to float is not supported
     [[Node: Adam/apply_grad_op_0/update_FullyConnected_1/b/Cast_2 = Cast[DstT=DT_FLOAT, SrcT=DT_STRING, _class=["loc:@FullyConnected_1/b"], _device="/job:localhost/replica:0/task:0/cpu:0"](Adam/apply_grad_op_0/learning_rate)]]
W tensorflow/core/framework/op_kernel.cc:983] Unimplemented: Cast string to float is not supported
E tensorflow/core/common_runtime/executor.cc:594] Executor failed to create kernel. Unimplemented: Cast string to float is not supported
     [[Node: Adam/apply_grad_op_0/update_FullyConnected_1/b/Cast_2 = Cast[DstT=DT_FLOAT, SrcT=DT_STRING, _class=["loc:@FullyConnected_1/b"], _device="/job:localhost/replica:0/task:0/cpu:0"](Adam/apply_grad_op_0/learning_rate)]]
W tensorflow/core/framework/op_kernel.cc:983] Unimplemented: Cast string to float is not supported
E tensorflow/core/common_runtime/executor.cc:594] Executor failed to create kernel. Unimplemented: Cast string to float is not supported
     [[Node: Adam/apply_grad_op_0/update_conv_1/W/Cast_2 = Cast[DstT=DT_FLOAT, SrcT=DT_STRING, _class=["loc:@conv_1/W"], _device="/job:localhost/replica:0/task:0/cpu:0"](Adam/apply_grad_op_0/learning_rate)]]

我正在使用tflearn.data_utils.image_preloader来读取png文件，但我也尝试使用其他一些方法，但似乎总是得到相同的错误。许多研究表明，这可能是由于图像文件不良造成的，但是在拍摄了十几个jpg图像之后我得到了同样的问题，所以它必须是其他的东西。任何建议都会受到赞赏，我的代码在下面，我的git here

上有更大的项目

import numpy as np
import tflearn
from tflearn.data_preprocessing import ImagePreprocessing
from tflearn.data_utils import image_preloader
from tflearn.layers.core import input_data, fully_connected, dropout
from tflearn.layers.conv import conv_2d, max_pool_2d
from tflearn.layers.estimator import regression


def train():

    training_path = 'images/train'

    image_height = 32
    image_width = 32
    colour_channels = 3

    X, Y = image_preloader(
        training_path,
        image_shape=(image_height, image_width),
        mode='folder',
        categorical_labels=True,
        normalize=True)

    X = np.reshape(X, (-1, image_height, image_width, colour_channels))

    img_prep = ImagePreprocessing()
    img_prep.add_featurewise_zero_center()
    img_prep.add_featurewise_stdnorm()

    network = input_data(shape=[None, image_height, image_width, colour_channels],
                         data_preprocessing=img_prep,
                         name='input')

    network = conv_2d(network, 32, 3, activation='relu', name='conv_1')
    network = max_pool_2d(network, 2)
    network = conv_2d(network, 64, 3, activation='relu', name='conv_2')
    network = conv_2d(network, 64, 3, activation='relu', name='conv_3')
    network = max_pool_2d(network, 2)
    network = fully_connected(network, 512, activation='relu')
    network = dropout(network, 0.5)
    network = fully_connected(network, 2, activation='softmax')

    network = regression(
        network,
        optimizer='adam',
        loss='categorical_crossentropy',
        learning_rate='0.001')

    model = tflearn.DNN(
        network,
        checkpoint_path='tmp/tflearn/cnn/checkpoints/model.tflearn',
        tensorboard_verbose=3,
        tensorboard_dir='tmp/tflearn/cnn/logs/')

    model.fit(
        X, Y,
        validation_set=0.2,
        n_epoch=1000,
        shuffle=True,
        batch_size=100,
        run_id='model',
        snapshot_epoch=True)

    model.save('tmp/tflearn/cnn/model/model_final.tflearn')

Answer 1

很抱歉迟到的回答（可能会帮助他人）

我遇到了同样的问题，问题出现在您编写的回归代码中

learning_rate='0.001'

但学习率为float而不是string，所以只需写一下：

learning_rate = 0.001

它会起作用

Tensorflow Executor无法创建内核。未实现：不支持将字符串转换为float

1 个答案: