astorfi / tensorflow-world Goto Github PK

View Code? Open in Web Editor NEW

4.5K 168.0 416.0 25.77 MB

:earth_americas: Simple and ready-to-use tutorials for TensorFlow

License: MIT License

Python 50.27% Shell 0.89% Jupyter Notebook 48.84%

deep-learning neural-network tensorflow machine-learning python computer-vision

tensorflow-world's Introduction

TensorFlow World

To support maintaining and upgrading this project, please kindly consider Sponsoring the project developer.

Any level of support is a great contribution here ❤️

This repository aims to provide simple and ready-to-use tutorials for TensorFlow. The explanations are present in the wiki associated with this repository.

Each tutorial includes source code and associated documentation.

Slack Group

Motivation

There are different motivations for this open source project. TensorFlow (as we write this document) is one of / the best deep learning frameworks available. The question that should be asked is why has this repository been created when there are so many other tutorials about TensorFlow available on the web?

Why use TensorFlow?

Deep Learning is in very high interest these days - there's a crucial need for rapid and optimized implementations of the algorithms and architectures. TensorFlow is designed to facilitate this goal.

The strong advantage of TensorFlow is it flexibility in designing highly modular models which can also be a disadvantage for beginners since a lot of the pieces must be considered together when creating the model.

This issue has been facilitated as well by developing high-level APIs such as Keras and Slim which abstract a lot of the pieces used in designing machine learning algorithms.

The interesting thing about TensorFlow is that it can be found anywhere these days. Lots of the researchers and developers are using it and its community is growing at the speed of light! So many issues can be dealt with easily since they're usually the same issues that a lot of other people run into considering the large number of people involved in the TensorFlow community.

What's the point of this repository?

Developing open source projects for the sake of just developing something is not the reason behind this effort. Considering the large number of tutorials that are being added to this large community, this repository has been created to break the jump-in and jump-out process that usually happens to most of the open source projects, but why and how?

First of all, what's the point of putting effort into something that most of the people won't stop by and take a look? What's the point of creating something that does not help anyone in the developers and researchers community? Why spend time for something that can easily be forgotten? But how we try to do it? Even up to this very moment there are countless tutorials on TensorFlow whether on the model design or TensorFlow workflow.

Most of them are too complicated or suffer from a lack of documentation. There are only a few available tutorials which are concise and well-structured and provide enough insight for their specific implemented models.

The goal of this project is to help the community with structured tutorials and simple and optimized code implementations to provide better insight about how to use TensorFlow quick and effectively.

It is worth noting that, the main goal of this project is to provide well-documented tutorials and less-complicated code!

TensorFlow Installation and Setup the Environment

In order to install TensorFlow please refer to the following link:

TensorFlow Installation

The virtual environment installation is recommended in order to prevent package conflict and having the capacity to customize the working environment.

TensorFlow Tutorials

The tutorials in this repository are partitioned into relevant categories.

Warm-up

#	topic	Source Code
1	Start-up	Welcome / IPython	Documentation

Basics

#	topic	Source Code
2	TensorFLow Basics	Basic Math Operations / IPython	Documentation
3	TensorFLow Basics	TensorFlow Variables / IPython	Documentation

Basic Machine Learning

#	topic	Source Code
4	Linear Models	Linear Regression / IPython	Documentation
5	Predictive Models	Logistic Regression / IPython	Documentation
6	Support Vector Machines	Linear SVM / IPython
7	Support Vector Machines	MultiClass Kernel SVM / IPython

Neural Networks

#	topic	Source Code
8	Multi Layer Perceptron	Simple Multi Layer Perceptron / IPython
9	Convolutional Neural Network	Simple Convolutional Neural Networks	Documentation
10	Autoencoder	Undercomplete Autoencoder	Documentation
11	Recurrent Neural Network	RNN / IPython

Some Useful Tutorials

TensorFlow Examples - TensorFlow tutorials and code examples for beginners

Sungjoon's TensorFlow-101 - TensorFlow tutorials written in Python with Jupyter Notebook

Terry Um’s TensorFlow Exercises - Re-create the codes from other TensorFlow examples

Classification on time series - Recurrent Neural Network classification in TensorFlow with LSTM on cellphone sensor data

Contributing

When contributing to this repository, please first discuss the change you wish to make via issue, email, or any other method with the owners of this repository before making a change. For typos, please do not create a pull request. Instead, declare them in issues or email the repository owner.

Please note we have a code of conduct, please follow it in all your interactions with the project.

Pull Request Process

Please consider the following criterions in order to help us in a better way:

The pull request is mainly expected to be a code script suggestion or improvement.

A pull request related to non-code-script sections is expected to make a significant difference in the documentation. Otherwise, it is expected to be announced in the issues section.

Ensure any install or build dependencies are removed before the end of the layer when doing a build and creating a pull request.

Add comments with details of changes to the interface, this includes new environment variables, exposed ports, useful file locations and container parameters.

You may merge the Pull Request in once you have the sign-off of at least one other developer, or if you do not have permission to do that, you may request the owner to merge it for you if you believe all checks are passed.

Final Note

We are looking forward to your kind feedback. Please help us to improve this open source project and make our work better. For contribution, please create a pull request and we will investigate it promptly. Once again, we appreciate your kind feedback and elaborate code inspections.

Acknowledgement

I have taken huge efforts in this project for hopefully being a small part of TensorFlow world. However, it would not have been plausible without the kind support and help of my friend and colleague Domenick Poster for his valuable advices. He helped me for having a better understanding of TensorFlow and my special appreciation goes to him.

tensorflow-world's People

Contributors

Stargazers

Watchers

Forkers

yaserkl bertomartin valeman hadikazemi little1tow mjunaidi ml-lab curtiszimmerman abhi-jha scholltan phillyc kaizen-soze vdt corefan nikolaevarius sevenyoung mwvaughn kidgodzilla seejay vamsijkrishna heyuhere dreadlord1984 anishsingh20 bodidze davidduan1989 ajnovice jacksonkearl lulzzz xuesj leon-root jai2033shankar vigyanik codernaut jdoliner kod3r tusharchhibber renatosc bigmaye sktgintokii practise2017 yanghaha11514 bborna bluenine99 mtfelix allensmile nicals gongqingyi-github jamiegoodwin2 ashu-22 tadasvas chiragsingla benjamesbabala austinpoon guptarohit thomasjfox1 tngamemo empia wenhel programmingtips gsjclgz pmmarq faisal-w gavinhzh wuyuzaizai hhy5277 vcalderon2009 fendouai liuyusg kitster deepak- cortez-zhang 554288653 lyrl hemangmehta dingshanliang hanuor ajmssc bechirghodbane ubaidsayyed54 sayantanmukherjee6 jmolinap yangchandle anandagarwaal-zz 19ai ashishchopra saeta yak0xff cuent qiuwen-sz roshanadk matrixy keylorsdu imvetri zgsxwsdxg ding-ye ayushyadav wxdublin shshao tangxinkevin enlyn

tensorflow-world's Issues

train_op in linear regression

Is defining train_op for each data point and epoch anew really needed? I'm new to TensorFlow so I can't tell why or why not this would make sense. For me, the regression seems to work fine (and much faster) if the line is removed.

RE: Policy regarding typos in codebase.

This issue is regarding your policy regarding typos in your codebase. Here is the relevant section in your CONTRIBUTING.rst:
For typos, please do not create a pull request. Instead, declare them in issues or email the repository owner.

I suggest this policy be revised as it creates an extra step for you, the maintainer of this repo. For example, here is your current process:

Contributor finds a typo.
Contributor opens an issue.
Repo owner reads the issue.
Repo owner decides to create a code change to fix the typo and pushes the change.

Here is the suggested process:

Contributor finds a typo.
Contributor creates a code change to fix the typo and creates a pull request
Repo owner decides to accept the pull request and merges the changes.

If typos can be discussed within a pull request, I don't see the point for a contributor to create an issue and then the repo owner creates a code change to fix the typo. I suggest using Github Issues to discuss lengthy proposals, but typos should be handled directly within a pull request. For example, see this Contributing guide for Github's open source guide.

sudo apt-get install nvidia-current-updates nvidia-settings-updates error

Hello,
just wanted to say this is a great guide.
but when i execute :
sudo apt-get install nvidia-current-updates nvidia-settings-updates
its says:
E: Unable to locate package nvidia-settings-updates

can someone help me with this?

TensorBoard

using tensorboard --logdir="absolute/path/to/log_dir" to run tensorboard seems to give an error. Removing the = solves the issue.

Typos in README.rst

There are several typing errors in the 'README.rst' file.

TensorFlow

$ git clone https://github.com/TensorFlow-World.git
Cloning into 'TensorFlow-World'...
remote: Not Found
fatal: repository 'https://github.com/TensorFlow-World.git/' not found

No Transformer Notebook

Hey,

I see that there are no tutorial notebooks for Transformer implementations in this repository yet. Transformers are used primarily in the field of natural language processing. Like recurrent neural networks, Transformers are designed to handle sequential data, such as natural language, for tasks such as translation and text summarization.

I would like to add such tutorial notebooks.

linear regression tutorial cost only reported for last data point

I noticed in the notebook for the linear regression that the cost was only being calculated for the last piece of data in each epoch.

with tf.Session() as sess:

    # Initialize the variables[w and b].
    sess.run(tf.global_variables_initializer())

    # Get the input tensors
    X, Y = inputs()

    # Return the train loss and create the train_op.
    train_loss = loss(X, Y)
    train_op = train(train_loss)

    # Step 8: train the model
    for epoch_num in range(num_epochs): # run 100 epochs
        for x, y in data:
          train_op = train(train_loss)

          # Session runs train_op to minimize loss
          loss_value,_ = sess.run([train_loss,train_op], feed_dict={X: x, Y: y})

        # Displaying the loss per epoch.
        print('epoch %d, loss=%f' %(epoch_num+1, loss_value))

        # save the values of weight and bias
        wcoeff, bias = sess.run([W, b])

data is being iterated over and the loss_value that is calculated is written over each time through the loop. Thus, the loss is only for the last piece of data. Since the loss needs to be computed over all of the data being used to train, the cost function should probably be something more like the following:

def loss(X, Y):
    '''
    compute the loss by comparing the predicted value to the actual label.
    :param X: The inputs.
    :param Y: The labels.
    :return: The loss over the samples.
    '''

    # Making the prediction.
    Y_predicted = inference(X)
    return tf.reduce_sum(tf.squared_difference(Y, Y_predicted))/(2*data.shape[0])

With this change above, the training section could be changed to the following (with the looping over data removed completely):

with tf.Session() as sess:

    # Initialize the variables[w and b].
    sess.run(tf.global_variables_initializer())

    # Get the input tensors
    X, Y = inputs()

    # Return the train loss and create the train_op.
    train_loss = loss(X, Y)
    train_op = train(loss(X, Y))

    # Step 8: train the model
    for epoch_num in range(num_epochs): # run 100 epochs
        loss_value, _ = sess.run([train_loss,train_op], feed_dict={X: data[:,0], Y: data[:,1]})

        # Displaying the loss per epoch.
        print('epoch %d, loss=%f' %(epoch_num+1, loss_value))

        # save the values of weight and bias
        wcoeff, bias = sess.run([W, b])

This would result in output like the following:

epoch 1, loss=1573.599976
epoch 2, loss=1332.513916
epoch 3, loss=1128.868408
epoch 4, loss=956.848999
epoch 5, loss=811.544067

I would be glad to submit a pull request with these and other minor changes. Please let me know if I have some misunderstanding.

logits is an undefined name in this context, should it be logits_last?

Undefined names can raise NameErrorat runtime.

https://travis-ci.org/astorfi/TensorFlow-World/jobs/272817788#L623-L626

https://github.com/astorfi/TensorFlow-World/blob/master/codes/3-neural_networks/multi-layer-perceptron/code/test_classifier.py#L113

a small mistake in doc

In the tutorial doc of chapter 1 "Basics/variables", there might be a misktake here:

# "variable_list_custom" is the list of variables that we want to initialize.
variable_list_custom = [weights, custom_variable]

# The initializer
init_custom_op = tf.variables_initializer(var_list=all_variables_list)

The last line of the code above might end up with var_list=variable_list_custom, not all_variables_list.

Here's url of the doc:
https://github.com/astorfi/TensorFlow-World/tree/master/docs/tutorials/1-basics/variables#initializing-specific-variables
Thank you for your repo, it helps me a lot.