Mansi145 / Data-Modeling

Data Modelling for Crop Analyzer

Dataset used : https://www.kaggle.com/emmarex/plantdisease

Tensorflow Installtion on Linux

Prerequisites

1. Python 3.3 or higher
2. At least 1GB of RAM

Step 1: Install python3-venv

apt-get install python3-venv -y

Step 2: Create and activate a Python virtual environment

python3 -m venv venv
source ./venv/bin/activate

output:

(venv) root@ubuntu:~#

Step 3: Update PIP

pip install -U pip

Step 4: Update setuptools

pip install -U setuptools

Step 5: Install TensorFlow

pip install tensorflow

Setting up Tensorflow on Windows

Tensorflow is a popular end-to-end Machine Learning Library from Google whcih supports developing Machine Learning Models and Data Modelling. To setup Tensorflow on Windows, you have to ensure that the necessary system requirements and prerequisties are followed:

• Python 3.5-3.7 should be installed on your System Path. You can check out your Python Version by opening the Command Prompt and typing in: python --version
• pip 19.0 or a higher version should be installed which enables to download Third Party Python Packages. You can check out your Python Version by opening the Command Prompt and typing in: pip --version
• Windows 7 or later (64-bit)
• GPU Support with a CUDA-enabled Card Once the necessary requirements and prerequisties have been verified, we will know move onto setting up Tensorflow on Windows by creating a Development Environment. For this purpose we are going to use virtualenv which is a tool to create isolated environments where we can install the necessary directories. To install a Virtual Environment, follow the following steps:

1. Open the Command Prompt.
2. Go to the C:\ directory in the Command Prompt.
3. Type the Command: pip3 install -U pip virtualenv
4. Check if Virtual Environment has been setup or not by running the following command: virtualenv --version

Since we have installed a Virtual Environment, we will now try to set up a Virtual Environment where we can install Tensorflow:

1. We will first create a Virtual Environment by making a ./venv directory: virtualenv --system-site-packages -p python3 ./venv
2. Now we will activate the Virtual Environment: .\venv\Scripts\activate
3. Now we will see what are the list of packages that have already been installed using pip: pip list
4. Incase the pip version has been depercated, it is advisable to follow these steps to upgrade the pip: pip install --upgrade pip
5. To install Tensorflow, now type in the command: pip install tensorflow
6. In case you want to exit the virtual environment you can just type in the command: deactivate

These commands will install Tensorflow on a Windows System with ease using pip. In case you are using an Anaconda Navigator, there are seperate set of commands which are explained here:

1. Open Command Prompt and type in the command: conda info This command will display all of the necessary details of Anaconda including the Active Environments and more.
2. Create an Environment on Conda using the following command: conda create -n [environment-name] Let's say I want to name my environment as "TensorTest" so my command would be: conda create -n tensortest
3. Now we will activate the command that we have just created: activate tensortest
4. Now run the following environment that you have just created: conda install tensorflow

Now we have seen how we can install Tensorflow on Windows using pip and conda. Now let's verify that we have installed Tensorflow successfully by running a simple Tensorflow Code:

1. Open the Command Prompt.
2. Type the following command to kickstart a Python Shell: python
3. Now type the following code:

>>> import tensorflow as tf
>>> output=tf.constant("Hello World")
>>> sess=tf.Session()
>>> print(sess.run(output))
'Hello World'
>>> sess.close()

We will see an output on the Command Prompt Screen which will verify that Tensorflow is now up and running on our system.