环境：ubuntu18 + nvidia 430 + cuda 10.0 + cudnn7.6.0 + tensorflow-gpu 2.0.0
调用 layers.Conv2D() 就报错，报错信息：
Epoch 1/5
2019-10-11 21:50:00.814925: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0
2019-10-11 21:50:01.026836: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2019-10-11 21:50:01.472727: E tensorflow/stream_executor/cuda/cuda_dnn.cc:329] Could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR
2019-10-11 21:50:01.476545: E tensorflow/stream_executor/cuda/cuda_dnn.cc:329] Could not create cudnn handle: CUDNN_STATUS_INTERNAL_ERROR
2019-10-11 21:50:01.476593: W tensorflow/core/common_runtime/base_collective_executor.cc:216] BaseCollectiveExecutor::StartAbort Unknown: 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 sequential/conv2d/Conv2D}}]]
64/60000 [..............................] - ETA: 16:19Traceback (most recent call last):
File "/media/dxs/E/Project/AI/PycharmProject/deepLearningProject-201903/tf2_demo/mnist_cnn.py", line 53, in
model.fit(train_images, train_labels, epochs=5,batch_size=64)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/training.py", line 728, in fit
use_multiprocessing=use_multiprocessing)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/training_v2.py", line 324, in fit
total_epochs=epochs)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/training_v2.py", line 123, in run_one_epoch
batch_outs = execution_function(iterator)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/keras/engine/training_v2_utils.py", line 86, in execution_function
distributed_function(input_fn))
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/def_function.py", line 457, in __call__
result = self._call(*args, **kwds)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/def_function.py", line 520, in _call
return self._stateless_fn(*args, **kwds)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1823, in __call__
return graph_function._filtered_call(args, kwargs) # pylint: disable=protected-access
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1141, in _filtered_call
self.captured_inputs)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1224, in _call_flat
ctx, args, cancellation_manager=cancellation_manager)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 511, in call
ctx=ctx)
File "/home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/eager/execute.py", line 67, in quick_execute
six.raise_from(core._status_to_exception(e.code, message), None)
File "", line 3, in raise_from
tensorflow.python.framework.errors_impl.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 sequential/conv2d/Conv2D (defined at home/dxs/anaconda3/envs/tf2/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py:1751) ]] [Op:__inference_distributed_function_1000]
Function call stack:
distributed_function
尝试过 升级cuda到10.1 还是报错，经过一番查找，发现只要在开头设置下 就可以了， 在开头加上以下语句：
# # =====================================================================
# # 不加这几句，则CONV 报错
# physical_devices = tf.config.experimental.list_physical_devices('GPU')
# assert len(physical_devices) > 0, "Not enough GPU hardware devices available"
# tf.config.experimental.set_memory_growth(physical_devices[0], True)
# # =========================================================================
发个完整代码：
# -*- coding: utf-8 -*-
# @Time : 2019/10/10 下午10:39
# @Author : dxs
# @Email : dangxusheng163163.com
# @File : mnist_cnn.py
# @Project : PycharmProject
from __future__ import absolute_import, division, print_function, unicode_literals
import os
import os.path as osp
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import datasets, layers, models
import cv2
import numpy as np
import matplotlib.pyplot as plt
#
# =====================================================================
# 不加这几句，则CONV 报错
physical_devices = tf.config.experimental.list_physical_devices('GPU')
assert len(physical_devices) > 0, "Not enough GPU hardware devices available"
tf.config.experimental.set_memory_growth(physical_devices[0], True)
# =========================================================================
(train_images, train_labels), (test_images, test_labels) = datasets.mnist.load_data()
train_images = train_images.reshape((60000, 28, 28, 1))
test_images = test_images.reshape((10000, 28, 28, 1))
# 特征缩放[0, 1]区间
train_images, test_images = train_images / 255.0, test_images / 255.0
model = models.Sequential()
model.add(layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.MaxPooling2D((2, 2)))
model.add(layers.Conv2D(64, (3, 3), activation='relu'))
model.add(layers.Flatten())
model.add(layers.Dense(64, activation='relu'))
model.add(layers.Dense(10, activation='softmax'))
model.summary() # 显示模型的架构
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
model.fit(train_images, train_labels, epochs=5,batch_size=64)
输出结果如下：
2019-10-11 21:54:15.109707: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-10-11 21:54:15.109714: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165] 0 2019-10-11 21:54:15.109717: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0: N 2019-10-11 21:54:15.109766: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1006] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2019-10-11 21:54:15.109991: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1006] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2019-10-11 21:54:15.110211: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1304] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5562 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1660, pci bus id: 0000:01:00.0, compute capability: 7.5) Model: "sequential" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= conv2d (Conv2D) (None, 26, 26, 32) 320 _________________________________________________________________ max_pooling2d (MaxPooling2D) (None, 13, 13, 32) 0 _________________________________________________________________ conv2d_1 (Conv2D) (None, 11, 11, 64) 18496 _________________________________________________________________ max_pooling2d_1 (MaxPooling2 (None, 5, 5, 64) 0 _________________________________________________________________ conv2d_2 (Conv2D) (None, 3, 3, 64) 36928 _________________________________________________________________ flatten (Flatten) (None, 576) 0 _________________________________________________________________ dense (Dense) (None, 64) 36928 _________________________________________________________________ dense_1 (Dense) (None, 10) 650 ================================================================= Total params: 93,322 Trainable params: 93,322 Non-trainable params: 0 _________________________________________________________________ Train on 60000 samples Epoch 1/5 2019-10-11 21:54:15.965416: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10.0 2019-10-11 21:54:16.172605: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7 2019-10-11 21:54:17.364824: W tensorflow/stream_executor/cuda/redzone_allocator.cc:312] Not found: ./bin/ptxas not found Relying on driver to perform ptx compilation. This message will be only logged once. 60000/60000 [==============================] - 9s 145us/sample - loss: 0.1734 - accuracy: 0.9475 Epoch 2/5 60000/60000 [==============================] - 8s 127us/sample - loss: 0.0504 - accuracy: 0.9844 Epoch 3/5 60000/60000 [==============================] - 7s 124us/sample - loss: 0.0365 - accuracy: 0.9886 Epoch 4/5 60000/60000 [==============================] - 7s 113us/sample - loss: 0.0290 - accuracy: 0.9902 Epoch 5/5 60000/60000 [==============================] - 7s 110us/sample - loss: 0.0229 - accuracy: 0.9929 Process finished with exit code 0