Numpy: This package is the basis for scientific computing with Python. It supports computing n-dimensional arrays and matrices, also we have a large number of advanced mathematical functions. This is necessary required packages TensorFlow, therefore, use pip install tensorflow, if not already installed Numpy, it will be automatically installed.
Matplolib: This is a Python 2D graphics library. Use it to create all kinds Figure only a few lines of code, including histograms, bar charts, error, scatter, and power spectrum. It can use pip to install:
OS: It provides a use operating system-related functions (such as read, write and change files and directories) simple portable way.
Pandas: This provides a variety of data structures and data analysis tools. Use Pandas, you can read and write data between the memory and data structures in different formats. .Csv files can be read and text. Pip install or can be installed conda install.
Keras: Advanced API Keras is the neural network, which uses TensorFlow as its backend. It can run on Theano and CNTK. Add a layer need only one line of code, very user friendly and can be installed using pip install keras.
TFlearn: TFlearn is to establish a modular and transparent depth study of the library on TensorFlow. It provides a higher level API for TensorFlow, to facilitate and speed up experiments. It currently supports the most recent model of deep learning, such as convolution, LSTM, BatchNorm, BiRNN, PReLU, residual network and build networks. It applies only to TensorFlow 1.0 or later. Please use pip install tflearn installation.
sklearn: Scikit-learn (sklearn) is commonly used in machine learning third party modules, machine learning methods commonly used the package, including regression (Regression), dimensionality reduction (Dimensionality Reduction), classification (Classfication), clustering (Clustering ) and other methods. Sklearn provides some standard data, we do not have to look for training data from other sites. Sklearn installation requirements Python (> = 2.7 or> = 3.3), NumPy (> = 1.8.2), SciPy (> = 0.13.3). If you have installed and NumPy SciPy, may be used to install scikit-learn pip install -U scikit-learn.
