我使用CUDA Toolkit 10.1 CUDNN 7.5.0（Windows 10）在tensorflow 2.3.0 Anaconda上运行我的代码，并返回问题
F .\tensorflow/core/kernels/random_op_gpu.h:246] Non-OK-status: GpuLaunchKernel(FillPhiloxRandomKernelLaunch<Distribution>, num_blocks, block_size, 0, d.stream(), key, counter, gen, data, size, dist) status: Internal: invalid configuration argument
我的GPU是RTX2070，我正在测试的代码是
import numpy as np
import os
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.models import Sequential
model = tf.keras.models.Sequential([
tf.keras.layers.Flatten(input_shape=(28, 28)),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dropout(0.2),
tf.keras.layers.Dense(10, activation='softmax')
])
model.compile(optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])
完整的结果在这里
2020-11-03 09:59:18.494825: I tensorflow/stream_executor/platform/default/dso_loader.cc:49]
Successfully opened dynamic library cudart64_110.dll
2020-11-03 09:59:20.388914: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set
2020-11-03 09:59:20.389652: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library nvcuda.dll
2020-11-03 09:59:20.426874: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1724] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce RTX 2070 computeCapability: 7.5
coreClock: 1.62GHz coreCount: 36 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 417.29GiB/s
2020-11-03 09:59:20.427039: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2020-11-03 09:59:20.435227: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2020-11-03 09:59:20.437546: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2020-11-03 09:59:20.448543: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2020-11-03 09:59:20.451378: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2020-11-03 09:59:20.464548: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2020-11-03 09:59:20.472311: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2020-11-03 09:59:20.506843: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2020-11-03 09:59:20.507014: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1866] Adding visible gpu devices: 0
2020-11-03 09:59:20.507910: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2020-11-03 09:59:20.508416: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1724] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce RTX 2070 computeCapability: 7.5
coreClock: 1.62GHz coreCount: 36 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 417.29GiB/s
2020-11-03 09:59:20.508536: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll
2020-11-03 09:59:20.508777: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll
2020-11-03 09:59:20.509056: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll
2020-11-03 09:59:20.509324: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll
2020-11-03 09:59:20.509572: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll
2020-11-03 09:59:20.509811: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll
2020-11-03 09:59:20.510030: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll
2020-11-03 09:59:20.510102: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll
2020-11-03 09:59:20.510384: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1866] Adding visible gpu devices: 0
2020-11-03 09:59:20.952560: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1265] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-03 09:59:20.952716: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1271] 0
2020-11-03 09:59:20.952746: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1284] 0: N
2020-11-03 09:59:20.953709: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1410] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6637 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2070, pci bus id: 0000:01:00.0, compute capability: 7.5)
2020-11-03 09:59:20.954420: I tensorflow/compiler/jit/xla_gpu_device.cc:99] Not creating XLA devices, tf_xla_enable_xla_devices not set
2020-11-03 09:59:21.179776: F .\tensorflow/core/kernels/random_op_gpu.h:246] Non-OK-status: GpuLaunchKernel(FillPhiloxRandomKernelLaunch<Distribution>, num_blocks, block_size, 0, d.stream(), key, counter, gen, data, size, dist) status: Internal: invalid configuration argument
之前有人遇到过吗？有什么问题以及如何解决？
我还测试了Tensorflow 2.1.0、2.2.0。发生了同样的问题...谢谢！
答案 0 :(得分：1)
自己想办法。当您忘记初始化 GPU 时就会出现这种情况。
添加以下代码解决问题
import tensorflow as tf
gpus = tf.config.experimental.list_physical_devices('GPU')
if gpus:
try:
# Currently, memory growth needs to be the same across GPUs
for gpu in gpus:
tf.config.experimental.set_memory_growth(gpu, True)
logical_gpus = tf.config.experimental.list_logical_devices('GPU')
print(len(gpus), "Physical GPUs,", len(logical_gpus), "Logical GPUs")
except RuntimeError as e:
# Memory growth must be set before GPUs have been initialized
print(e)
答案 1 :(得分：-1)
也许你的 CUDA cdunn 和 tensorflow 版本不对
