Keras代码在Jupyter中不起作用：“内核似乎已经死亡。它将自动重新启动。”

我正在用Keras编写代码，用于基于深度学习的简单30x30猫图像分类器。当我到达应该训练模型的代码部分时，Jupyter停止运行，并显示错误消息“内核似乎已经死亡。它将自动重新启动。”我不知道是什么原因导致了这种情况的发生。如果在终端中查看，我将得到CUDA_ERROR_OUT_OF_MEMORY：内存不足；报告的总内存：11520114688。我认为一个简单的分类器不会耗尽PC的资源。我有RTX 2080TI，32GB内存，i9-9900k。

我不知道这是软件的兼容性问题还是什么。但是我确实知道tensorflow-gpu正在工作，因为在我的控制台中它是如此。我使用的代码本质上是逐字逐句地从Keras书中的深度学习中获得的。该代码在我6岁的笔记本电脑上运行良好，尽管训练得很慢。

```from keras import layers
```from keras import models
-------------
```model = models.Sequential()
```model.add(layers.Conv2D(32,(3,3),activation='relu',input_shape=(150,150,3)))
```model.add(layers.MaxPooling2D((2,2)))
```model.add(layers.Conv2D(64,activation='relu'))
```model.add(layers.MaxPooling2D((2,2)))
```model.add(layers.Conv2D(128,2)))
```model.add(layers.flatten())
```model.add(layers.Dropout(0.5))
```model.add(layers.Dense(512,activation='relu'))
```model.add(layers.Dense(1,activation='sigmoid'))

在上面运行此块之后：警告：tensorflow：来自/home/name/venv/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263：colocate_with（来自tensorflow.python.framework.ops）已被弃用在将来的版本中删除。更新说明：托管由托管服务器自动处理。警告：tensorflow：从/home/name/venv/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445：不推荐使用keep_prob调用dropout（来自tensorflow.python.ops.nn_ops）在将来的版本中删除。更新说明：

请使用`rate`代替`keep_prob`。费率应设置为`rate = 1 - keep_prob`。

```from keras import optimizers
```model.compile(loss='binary_crossentropy',optimizer=optimizers.RMSprop(lr=1e-4),metrics=['acc'])
```from keras.preprocessing.image import ImageDataGenerator
```train_datagen = ImageDataGenerator(rescale=1./255,rotation_range=40,width_shift_range=0.2,height_shift_range=0.2,shear_range=0.2,zoom_range=0.2,horizontal_flip=True)
```test_datagen = ImageDataGenerator(rescale=1./255)

```import os,shutil
```base_dir = '/home/name/Desktop/Deep Learning/Cat-Dog ```exercise/cats_and_dogs_small'
```train_dir = os.path.join(base_dir,'train')
```validation_dir = os.path.join(base_dir,'validation')
```test_dir = os.path.join(base_dir,'test')

```train_generator = train_datagen.flow_from_directory(train_dir,target_size=(150,150),batch_size=10,class_mode='binary')
```validation_generator = test_datagen.flow_from_directory(validation_dir,target_size = (150,class_mode='binary')
```history = model.fit_generator(train_generator,steps_per_epoch=10,epochs=100,validation_data=validation_generator,validation_steps=50)
```model.save('cats_and_dogs_small_2.h5')

运行此命令后，这是输出：

找到2000个属于2类的图像。找到属于2类的1000张图像。警告：tensorflow：来自/home/name/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066：to_int32（来自tensorflow.python.ops.math_ops）已被弃用在将来的版本中删除。更新说明：请改用tf.cast。时代1/100-然后在这里崩溃，我收到了jupyter错误。

来自终端的其他复制粘贴：

[W 19:44:35.199 NotebookApp] Notebook Desktop/Deep Learning/TF test.ipynb is not trusted
[I 19:44:35.307 NotebookApp] Kernel started: ae9c0530-bdbb-4748-a8b1-4e9a98fad3b8
[I 19:44:35.669 NotebookApp] Adapting to protocol v5.1 for kernel ae9c0530-bdbb-4748-a8b1-4e9a98fad3b8
2019-11-11 19:44:38.840439: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 fma
2019-11-11 19:44:38.956151: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1),but there must be at least one NUMA node,so returning NUMA node zero
2019-11-11 19:44:38.957106: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x28f53e0 executing computations on platform CUDA. Devices:
2019-11-11 19:44:38.957121: I tensorflow/compiler/xla/service/service.cc:158]   StreamExecutor device (0): GeForce RTX 2080 Ti,Compute Capability 7.5
2019-11-11 19:44:38.984940: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3600000000 Hz
2019-11-11 19:44:38.985600: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x2953090 executing computations on platform Host. Devices:
2019-11-11 19:44:38.985611: I tensorflow/compiler/xla/service/service.cc:158]   StreamExecutor device (0): <undefined>,<undefined>
2019-11-11 19:44:38.986045: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1433] Found device 0 with properties: 
name: GeForce RTX 2080 Ti major: 7 minor: 5 memoryClockRate(GHz): 1.635
pciBusID: 0000:01:00.0
totalMemory: 10.73GiB freeMemory: 10.18GiB
2019-11-11 19:44:38.986059: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0
2019-11-11 19:44:38.987129: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-11-11 19:44:38.987140: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990]      0 
2019-11-11 19:44:38.987144: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0:   N 
2019-11-11 19:44:38.987375: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/device:GPU:0 with 9903 MB memory) -> physical GPU (device: 0,name: GeForce RTX 2080 Ti,pci bus id: 0000:01:00.0,compute capability: 7.5)
[I 19:44:43.456 NotebookApp] Saving file at /Desktop/Deep Learning/TF test.ipynb
[I 19:44:43.565 NotebookApp] Starting buffering for ae9c0530-bdbb-4748-a8b1-4e9a98fad3b8:a4ca004462174782820be082f11422c7
[W 19:44:46.645 NotebookApp] Notebook Desktop/Deep Learning/5-2 Cat,Dog.ipynb is not trusted
[I 19:44:46.838 NotebookApp] Kernel started: 38279d87-721e-44c2-a40e-48cd4c7ba1c4
[I 19:44:47.142 NotebookApp] Adapting to protocol v5.1 for kernel 38279d87-721e-44c2-a40e-48cd4c7ba1c4
2019-11-11 19:45:18.007748: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 fma
2019-11-11 19:45:18.095869: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1),so returning NUMA node zero
2019-11-11 19:45:18.096371: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x37ab830 executing computations on platform CUDA. Devices:
2019-11-11 19:45:18.096384: I tensorflow/compiler/xla/service/service.cc:158]   StreamExecutor device (0): GeForce RTX 2080 Ti,Compute Capability 7.5
2019-11-11 19:45:18.116837: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3600000000 Hz
2019-11-11 19:45:18.117825: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x2b14f90 executing computations on platform Host. Devices:
2019-11-11 19:45:18.117837: I tensorflow/compiler/xla/service/service.cc:158]   StreamExecutor device (0): <undefined>,<undefined>
2019-11-11 19:45:18.118228: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1433] Found device 0 with properties: 
name: GeForce RTX 2080 Ti major: 7 minor: 5 memoryClockRate(GHz): 1.635
pciBusID: 0000:01:00.0
totalMemory: 10.73GiB freeMemory: 10.03GiB
2019-11-11 19:45:18.118258: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0
2019-11-11 19:45:18.118879: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-11-11 19:45:18.118886: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990]      0 
2019-11-11 19:45:18.118907: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0:   N 
2019-11-11 19:45:18.119131: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9756 MB memory) -> physical GPU (device: 0,compute capability: 7.5)
2019-11-11 19:45:19.008213: I tensorflow/stream_executor/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally
[I 19:46:32.074 NotebookApp] Kernel interrupted: 38279d87-721e-44c2-a40e-48cd4c7ba1c4
[I 19:46:36.231 NotebookApp] Saving file at /Desktop/Deep Learning/5-2 Cat,Dog.ipynb
[W 19:46:36.232 NotebookApp] Notebook Desktop/Deep Learning/5-2 Cat,Dog.ipynb is not trusted
[I 19:46:36.348 NotebookApp] Starting buffering for 38279d87-721e-44c2-a40e-48cd4c7ba1c4:de4e1e737fef45e0a9393696f2ee53b4
[I 19:46:39.645 NotebookApp] Kernel started: 6192e4fe-6840-45b0-9f26-ffec2efcd443
[I 19:46:39.902 NotebookApp] Adapting to protocol v5.1 for kernel 6192e4fe-6840-45b0-9f26-ffec2efcd443
2019-11-11 19:46:53.523436: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 fma
2019-11-11 19:46:53.634301: W tensorflow/compiler/xla/service/platform_util.cc:240] unable to create StreamExecutor for CUDA:0: failed initializing StreamExecutor for CUDA device ordinal 0: Internal: failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_OUT_OF_MEMORY: out of memory; total memory reported: 11520114688
2019-11-11 19:46:53.634372: F tensorflow/stream_executor/lib/statusor.cc:34] Attempting to fetch value instead of handling error Internal: no supported devices found for platform CUDA
[I 19:46:54.645 NotebookApp] KernelRestarter: restarting kernel (1/5),keep random ports
WARNING:root:kernel 6192e4fe-6840-45b0-9f26-ffec2efcd443 restarted
[I 19:48:39.650 NotebookApp] Saving file at /Desktop/Deep Learning/5-2 cats,dogs w regularization.ipynb
2019-11-11 19:49:31.150869: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 fma
2019-11-11 19:49:31.223933: W tensorflow/compiler/xla/service/platform_util.cc:240] unable to create StreamExecutor for CUDA:0: failed initializing StreamExecutor for CUDA device ordinal 0: Internal: failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_OUT_OF_MEMORY: out of memory; total memory reported: 11520114688
2019-11-11 19:49:31.224025: F tensorflow/stream_executor/lib/statusor.cc:34] Attempting to fetch value instead of handling error Internal: no supported devices found for platform CUDA
[I 19:49:33.653 NotebookApp] KernelRestarter: restarting kernel (1/5),keep random ports
WARNING:root:kernel 6192e4fe-6840-45b0-9f26-ffec2efcd443 restarted
2019-11-11 19:49:51.505573: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 fma
2019-11-11 19:49:51.579785: W tensorflow/compiler/xla/service/platform_util.cc:240] unable to create StreamExecutor for CUDA:0: failed initializing StreamExecutor for CUDA device ordinal 0: Internal: failed call to cuDevicePrimaryCtxRetain: CUDA_ERROR_OUT_OF_MEMORY: out of memory; total memory reported: 11520114688
2019-11-11 19:49:51.579900: F tensorflow/stream_executor/lib/statusor.cc:34] Attempting to fetch value instead of handling error Internal: no supported devices found for platform CUDA
[I 19:49:51.665 NotebookApp] KernelRestarter: restarting kernel (1/5),keep random ports
WARNING:root:kernel 6192e4fe-6840-45b0-9f26-ffec2efcd443 restarted
[I 19:50:39.647 NotebookApp] Saving file at /Desktop/Deep Learning/5-2 cats,dogs w regularization.ipynb
[E 19:58:52.375 NotebookApp] nbconvert failed: No template_file specified!
    Traceback (most recent call last):
      File "/home/name/venv/lib/python3.6/site-packages/notebook/nbconvert/handlers.py",line 130,in get
        resources=resource_dict
      File "/home/name/venv/lib/python3.6/site-packages/nbconvert/exporters/templateexporter.py",line 315,in from_notebook_node
        output = self.template.render(nb=nb_copy,resources=resources)
      File "/home/name/venv/lib/python3.6/site-packages/nbconvert/exporters/templateexporter.py",line 113,in template
        self._template_cached = self._load_template()
      File "/home/name/venv/lib/python3.6/site-packages/nbconvert/exporters/templateexporter.py",line 278,in _load_template
        raise ValueError("No template_file specified!")
    ValueError: No template_file specified!
[W 19:58:52.376 NotebookApp] 500 GET /nbconvert/custom/Desktop/Deep%20Learning/5-2%20cats%2Cdogs%20w%20regularization.ipynb?download=true (127.0.0.1): nbconvert failed: No template_file specified!
[E 19:58:52.379 NotebookApp] {
      "Host": "localhost:8888","User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:70.0) Gecko/20100101 Firefox/70.0","accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8","accept-Language": "en-US,en;q=0.5","accept-Encoding": "gzip,deflate","Connection": "keep-alive","Referer": "http://localhost:8888/notebooks/Desktop/Deep%20Learning/5-2%20cats%2Cdogs%20w%20regularization.ipynb","Cookie": "_xsrf=2|79acfc45|e59d0dfc567c5357dd03be1910afbf02|1572817491; username-localhost-8888=\"2|1:0|10:1573523050|23:username-localhost-8888|44:N2U5M2FmY2QzYTE1NDBhZDg4NGY3N2U2MWU4MjYwOTU=|9895ebebc839d5dc7ba5a25aa7fb16775d32c546b63da08b9acd1c060bac0dde\"","Upgrade-Insecure-Requests": "1"

+ ---------------------------------------------- ------------------------------- + | NVIDIA-SMI 418.56驱动程序版本：418.56 CUDA版本：10.1 | | ------------------------------- + ----------------- ----- + ---------------------- + | GPU名称持久性-M |总线编号Disp.A |挥发性不佳。 ECC | |风扇温度性能：用法/上限|内存使用| GPU实用计算M。 | ============================== + ================= ===== + ==================== | | 0 GeForce RTX 208 ...关闭| 00000000：01：00.0开| N / A | | 0％52C P8 34W / 260W | 428MiB / 10986MiB | 2％违约| + ------------------------------- + ----------------- ----- + ---------------------- +

+ ---------------------------------------------- ------------------------------- + |进程：GPU内存| | GPU PID类型进程名称用法| | ================================================= ========================== | | 0 1253 G / usr / lib / xorg / Xorg 26MiB | | 0 1364 G / usr / bin / gnome-shell 58MiB | | 0 3027 G / usr / lib / xorg / Xorg 188MiB | | 0 3164 G / usr / bin / gnome-shell 139MiB | | 0 7687 G / usr / lib / firefox / firefox 6MiB | | 0 8085 G / usr / lib / firefox / firefox 6MiB | + ------------------------------------------------- ---------------------------- +

打包版本

absl-py 0.7.0
阿斯特0.7.1
attrs 18.2.0
回叫0.1.0
漂白剂3.1.0
循环仪0.10.0
装饰器4.3.2
defusedxml 0.5.0
入口点0.3
加气0.2.2
grpcio 1.19.0
h5py 2.9.0
ipykernel 5.1.0
ipython 7.3.0
ipython-genutils 0.2.0
ipywidgets 7.4.2
绝地0.13.3
Jinja2 2.10
jsonschema 3.0.0
jupyter 1.0.0
jupyter-client 5.2.4
jupyter-console 6.0.0
jupyter-core 4.4.0
Keras 2.2.4
Keras-应用程序1.0.7
Keras预处理1.0.9
kiwisolver 1.0.1
Markdown 3.0.1
MarkupSafe 1.1.1
matplotlib 3.0.2
误调0.8.4
模拟2.0.0
nbconvert 5.4.1
nbformat 4.4.0
笔记本5.7.4
numpy 1.16.2
熊猫0.24.1
pandocfilters 1.4.2
parso 0.3.4
pbr 5.1.3
pexpect 4.6.0
pickleshare 0.7.5
枕头5.4.1
点19.3.1
pkg-resources 0.0.0
prometheus-client 0.6.0
提示工具包2.0.9
protobuf 3.7.0
ptyprocess 0.6.0
Pygments 2.3.1
pyparsing 2.3.1
耐热0.14.11
python-dateutil 2.8.0
pytz 2018.9
PyYAML 3.13
pyzmq 18.0.0
qtconsole 4.4.3
scikit学习0.20.2
scipy 1.2.1
Send2Trash 1.5.0
setuptools 40.8.0
六个1.12.0
张量板1.13.0
tensorflow-estimator 1.13.0
tensorflow-gpu 1.13.1
termcolor 1.1.0
总站0.8.1
测试路径0.4.2
火炬1.0.1.post2 火炬视觉0.2.2.post2 龙卷风5.1.1
特质4.3.2
wcwidth 0.1.7
webencodings 0.5.1
Werkzeug 0.14.1
轮0.33.1
小部件nbextension 3.4.2

Keras代码在Jupyter中不起作用：“内核似乎已经死亡。它将自动重新启动。”

请使用rate代替keep_prob。费率应设置为rate = 1 - keep_prob。

a2271487 回答：Keras代码在Jupyter中不起作用：“内核似乎已经死亡。它将自动重新启动。”

大家都在问

请使用`rate`代替`keep_prob`。费率应设置为`rate = 1 - keep_prob`。