Note: If the installation is successful, the Anaconda version number is displayed. If an exception occurs, reinstall the Anaconda
Step2 install jupyter notebook and tensorflow
2.1 Create an independent Python-v3.6 environment based on Anaconda
Currently python37 support for tensorflow is not particularly friendly, and python3.6 is recommended
Run the command conda create -n tf-20190930 python=3.6
Tf-20190930: This is a custom name
After the python runtime environment is created, you can view the python runtime environment using the following command:conda info --env
2.2 Installing the IPykernel of Jupyter in an independent Python Environment
Before installing Jupyter, install ipykernel, which is the kernel of Jupyter
To activate a standalone Python environment:conda activate tf-20190930
Install ipykernel:conda install ipykernel
2.3 installation jupyter
Execute the instructions to install JUPyter:conda install nb_conda
(Very important) Write ipykernel to Jupyte and execute the command:python -m ipykernel install --user --name tf-20190930 --display-name 'tf-20190930'
parameter
describe
–user
User name (TF-20190930)
–name
The name of the Python standalone environment (TF-20190930)
–display-name
Represents the display name in Jupyter NoteBook (TF-20190930)
2.4 Installing KERas (including VGG16 and VGG19)
If you cannot download it directly, you need to configure the PIP source. Here we configure it as ali mirror source
PIP configuration files under Windows are located at:C:\Users\To Kill a MockinBird\pip.ini(Note: If the file does not exist, create it in this path.)
PIP configuration files under Linux are located at:/root/.pip/pip.conf
2.5 Two Methods of Starting Jupyter
The first method is to start from the Anaconda GUI
The second way is to start from the CMD command line
Start-up success
Step3 “VGG model “and” item classification table (imagenet_class_index.json) “are stored in the specified directory
3.1 Default storage path of pretraining model
Windows10 under the default path:C:\Users\To Kill a MockinBird\.keras\models;
Default path for Linux:/root/.keras/
3.2 Jupyter Create a Python program to invoke the VGG model through the API
3.3 Call VGG model to classify image types, the code is as follows:
# load model = VGG16 VGG16 model () from keras. Preprocessing, image import load_img # resource loading images, image storage paths by default: C:\Users\To Kill a MockinBird\dots.jpg image = load_img('dots.jpg', target_size=(224, 224) from keras. Preprocessing. Image import img_to_array # converts image pixel numpy array image = img_to_array image = (image) image.reshape((1, image.shape[0], image.shape[1], image.shape[2])) from keras.applications.vgg16 import preprocess_input # prepare the image for the VGG model image = preprocess_input(image) # predict the probability across all output classes yhat = model.predict(image) from keras.applications.vgg16 import decode_predictions # convert the probabilities to class labels label = decode_predictions(yhat) # retrieve the most likely result, Highest probability label = label[0][0] # print the classification print(' %s (%.2f%%)' % (label[1], label[2]*100))Copy the code
3.4 Click “Run” to perform VGG convolutional neural image classification
(Note: the default path of image storage is: C:\Users\To Kill a MockinBird\ dot.jpg)