神经网络 Python 错误“无法获得卷积算法”答案

【问题标题】：Neural Network Python Error "Failed to get convolution algorithm"神经网络 Python 错误“无法获得卷积算法”
【发布时间】：2019-12-06 14:19:04
【问题描述】：

我正在尝试在 Python 中运行一些神经网络代码。我在 Google Colab 上运行良好。然后，我将代码移至远程机器 GPU 上的 Jupyter Notebook。

在我尝试使用以下方法拟合模型之前，它运行正常：

history = model.fit_generator(generator=training_generator, validation_data=validation_generator, use_multiprocessing=True, workers=1, epochs=100, shuffle=True, verbose=1)

完整的错误信息如下。我只是不知道从哪里开始理解它的含义，所以我正在寻求帮助。提前致谢：

UnknownError                              Traceback (most recent call last)
<ipython-input-15-d3d33225fec8> in <module>
      1 # Train model on dataset
----> 2 history = model.fit_generator(generator=training_generator, validation_data=validation_generator, use_multiprocessing=True, workers=1, epochs=100, shuffle=True, verbose=1)

~/miniconda3/lib/python3.7/site-packages/keras/legacy/interfaces.py in wrapper(*args, **kwargs)
     89                 warnings.warn('Update your `' + object_name + '` call to the ' +
     90                               'Keras 2 API: ' + signature, stacklevel=2)
---> 91             return func(*args, **kwargs)
     92         wrapper._original_function = func
     93         return wrapper

~/miniconda3/lib/python3.7/site-packages/keras/engine/training.py in fit_generator(self, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
   1416             use_multiprocessing=use_multiprocessing,
   1417             shuffle=shuffle,
-> 1418             initial_epoch=initial_epoch)
   1419 
   1420     @interfaces.legacy_generator_methods_support

~/miniconda3/lib/python3.7/site-packages/keras/engine/training_generator.py in fit_generator(model, generator, steps_per_epoch, epochs, verbose, callbacks, validation_data, validation_steps, class_weight, max_queue_size, workers, use_multiprocessing, shuffle, initial_epoch)
    215                 outs = model.train_on_batch(x, y,
    216                                             sample_weight=sample_weight,
--> 217                                             class_weight=class_weight)
    218 
    219                 outs = to_list(outs)

~/miniconda3/lib/python3.7/site-packages/keras/engine/training.py in train_on_batch(self, x, y, sample_weight, class_weight)
   1215             ins = x + y + sample_weights
   1216         self._make_train_function()
-> 1217         outputs = self.train_function(ins)
   1218         return unpack_singleton(outputs)
   1219 

~/miniconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py in __call__(self, inputs)
   2713                 return self._legacy_call(inputs)
   2714 
-> 2715             return self._call(inputs)
   2716         else:
   2717             if py_any(is_tensor(x) for x in inputs):

~/miniconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py in _call(self, inputs)
   2673             fetched = self._callable_fn(*array_vals, run_metadata=self.run_metadata)
   2674         else:
-> 2675             fetched = self._callable_fn(*array_vals)
   2676         return fetched[:len(self.outputs)]
   2677 

~/miniconda3/lib/python3.7/site-packages/tensorflow/python/client/session.py in __call__(self, *args, **kwargs)
   1437           ret = tf_session.TF_SessionRunCallable(
   1438               self._session._session, self._handle, args, status,
-> 1439               run_metadata_ptr)
   1440         if run_metadata:
   1441           proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

~/miniconda3/lib/python3.7/site-packages/tensorflow/python/framework/errors_impl.py in __exit__(self, type_arg, value_arg, traceback_arg)
    526             None, None,
    527             compat.as_text(c_api.TF_Message(self.status.status)),
--> 528             c_api.TF_GetCode(self.status.status))
    529     # Delete the underlying status object from memory otherwise it stays alive
    530     # as there is a reference to status from this from the traceback due to

UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
     [[{{node conv2d_1/convolution}}]]
     [[{{node metrics/acc/Mean}}]]

【问题讨论】：

每当 CUDNN 与 CUDA 版本不兼容时，我都会看到此错误。你检查过你的 CUDNN、CUDA、Driver 是否兼容？
我不知道该怎么做 - 但感谢你给了我一个开始的地方:)
这个页面可能有用 tensorflow.org/install/gpu 。但可能涉及一些试验和错误

标签： python tensorflow keras neural-network

【解决方案1】：

正如@thushv89 所说，这是 TF 二进制文件和安装的 CUDNN 版本的兼容性问题。

您可以使用以下方法检查您的 tensorflow 版本：

python -c 'import tensorflow as tf; print(tf.__version__);'

在此处检查所需的 CUDA/CUDNN 版本： https://www.tensorflow.org/install/source#tested_build_configurations

注意： 表示 CUDA/CUDNN 版本仅与 TF 的官方分发相关。对于 conda 应该有更好的处理方式。

然后你可以检查你的CUDA版本：

nvcc --version

然后使用以下方法之一检查您的 CUDNN 版本：

cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
cat /usr/include/cudnn.h | grep CUDNN_MAJOR -A 2

【讨论】：