环境搭建

1、安装Python3.9.7。

2、修改pip源。在C:\Users\{你的用户名}\目录下创建一个名称是pip的文件夹，然后在文件夹中创建一个名为pip.ini的文件，在里面输入以下内容。

[global]
index-url=http://mirrors.aliyun.com/pypi/simple/

[install]
trusted-host=mirrors.aliyun.com

3、升级pip。

pip install --user --upgrade pip

4、安装vc_redist。

下载地址：https://docs.microsoft.com/zh-CN/cpp/windows/latest-supported-vc-redist?view=msvc-170

5、安装tensorflow。

pip install --user --upgrade tensorflow

6、验证安装效果。

python -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"

7、安装显卡驱动。

CUDA® 11.2 要求 450.80.02 或更高版本。

下载地址：https://www.nvidia.com/download/index.aspx?lang=en-us

8、安装CUDA® 11.2。（需要安装附带的CUPTI组件）

下载地址：https://developer.nvidia.com/cuda-toolkit-archive

9、安装cuDNN SDK 8.1.0。

下载地址：https://developer.nvidia.com/cudnn

将cuDNN解压到C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.2文件夹。

10、运行mnist_convnet.py查看结果。

2022-03-20 19:59:59.905384: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX AVX2

To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.

2022-03-20 20:00:00.481565: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 1673 MB memory:  -> device: 0, name: NVIDIA GeForce RTX 3050 Laptop GPU, pci bus id: 0000:01:00.0, compute capability: 8.6

Epoch 1/5

2022-03-20 20:00:02.261836: I tensorflow/stream_executor/cuda/cuda_dnn.cc:368] Loaded cuDNN version 8100

2022-03-20 20:00:04.870372: I tensorflow/stream_executor/cuda/cuda_blas.cc:1786] TensorFloat-32 will be used for the matrix multiplication. This will only be logged once.

938/938 [==============================] - 11s 6ms/step - loss: 0.1746 - accuracy: 0.9453

Epoch 2/5

938/938 [==============================] - 5s 6ms/step - loss: 0.0474 - accuracy: 0.9851

Epoch 3/5

938/938 [==============================] - 6s 6ms/step - loss: 0.0345 - accuracy: 0.9893

Epoch 4/5

938/938 [==============================] - 6s 6ms/step - loss: 0.0252 - accuracy: 0.9920

Epoch 5/5

938/938 [==============================] - 6s 6ms/step - loss: 0.0190 - accuracy: 0.9939

313/313 [==============================] - 1s 2ms/step - loss: 0.0275 - accuracy: 0.9912