Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
Sanjiban Sengupta
Hello everyone, Myself Sanjiban Sengupta, a junior from IIIT Bhubaneswar. I like to work in the fields of Machine learning, Deep learning, and previously used Tensorflow a lot in building and deploying models in various projects. I am interested to contribute to Tensorflow for GSoC'2021, looking forward to learn from you all.
Angad Bajwa
Hey everyone! I'm Angad Ripudaman Singh Bajwa, third year at NIT- Trichy. I'm interested in the field of Deep Learning, Computer Vision, and NLP. Really excited to be part of this community and to contribute to Tensorflow for Gsoc'21. I have read the instructions for contributing. Looking forward to learning and working with you all!!
Ansh Sarkar
Hello everyone! I am Ansh Sarkar, a junior at PES University pursuing B.Tech in Computer Science & Engineering. I have a keen interest in Machine learning and Deep learning. I have used TensorFlow extensively in different projects. I am comfortable with C, C++, and Python. I have also cleared the Google Tensorflow examination and have some experience in open source contributions. I am a GSoC'21 aspirant and would love to start contributing to Tensorflow. It will be really helpful if someone could help me out and get me started with the issues and contributions. I am really excited to get started and learning from you all!
Hello everyone, I'm Pradhi Das, third year at NIT Bhopal. I find the machine learning and AI domain very interesting and hence I am here to contribute to Tensorflow. May I know where and how can start contributing?
Mihai Maruseac

For all "I want to GSOC'21, help me start contributing" posts:

  1. Go to the TF repo
  2. Find issue
  3. Solve issue
  4. Go to step 2.
  5. Success

It's as simple as this

repeating this as it gets drowned by introductions

Alex Nguyen
but we stuck at step2 by infinite loop, we can never get to step 5 =))
Harshith MohanKumar

@rxng8 Most issues seem intimidating at first but after spending time carefully reading through the issue as well as the source code related to the issue you will get a better understanding of the problem at hand. Most people give up immediately after looking at a hard question. The key is persistence! Try to filter out issues that are beyond your understanding.

The best place to start is by solving documentation issues!

Alex Nguyen
Thank you for a great advice. I was just kidding ^^ But i really appreciate!
Harshith MohanKumar
Hopefully, it helps others who are stuck :P
Mihai Maruseac
doing a documentation PR is a good idea but keep in mind:
  • don't just update a single typo, it's better to take a file/directory and fix as much as possible from there
  • there is no GSoC project to only update documentation
I recommend doing a few documentation PRs to get accustomed with the flow but then switching to issues/features
I am looking for tf.image module on the Github repo of tensor flow, but unable to find it. Please guide me or provide some link.
Mihai Maruseac
just visit the page for the corresponding API and there's a "View source on GitHub" button (e.g., https://www.tensorflow.org/api_docs/python/tf/image/adjust_brightness)
or on Github/clone of repo search for tf.image
Vo Van Nghia

For all contributors who are struggling, here is my thought, hope it help !

TensorFlow are huge, doing a PR for Tensorflow takes a lot of time and effort. Things like import/copybara, feedback/copybaraand internal stuffs make TensorFlow not really friendly for external contributors. However, TensorFlow still has 93 other repositories ! Many repositories ( e.g https://github.com/tensorflow/io ) are fully external and their codebase are much smaller compared to TensorFlow. Contributions to these repositories are still counted as contributions to TensorFlow. Many GSoC projects in the past years were mainly targeted these repos.

Just discover a bit more and you will find !

@vnvo2409 thanks we will surely do it!!
Suraj Upadhyay
Assuming this is the right place to ask this (please prompt if it's not so), how may I make the tensorflow build process faster ? I mean I had built Linux kernel on the same machine and it took 4 times lesser time than this. Can I alot more threads to the build process ? How do I determine the progress of my build ? Is there any cap to the number [XXX / XXXX] which is shown while compile time ? Thanks in advance 😊
Mihai Maruseac
we are aware it takes a long time to compile; we have a plan to reduce the build time by some percents over the course of the year
but until then, the best you can do is watch to see if you're CPU bound or memory bound. If the build asks for too much memory and it swaps then you lose time during the swapping, can pass some Bazel filters to limit amount of jobs/memory used
if you are CPU bound, then adding more cores would help, if possible
then, using a remote cache (RBE) can also help with rebuilds
Suraj Upadhyay
Should submitting RFCs be accompanied by WIP pull requests showcasing the code ?
Mihai Maruseac
would be nice
or at least an API design should show up in the RFC
Suraj Upadhyay
Hii, where can I find design document for Tensorflow ? I just want to make a sense of the codebase
Aditya Nikhil
@SurajUpadhyay0_twitter There's no seperate design document I I think. It depends on, to which domain of tensorflow you want to contribute to.(there are many)
I suggest checking out their previous year ideas list. Check each domain and make a sense of contributing the code
I am interested in solving TensorFlow issues, I wonder where to seek help if I want to get started.
Atreya Majumdar
Hey everyone! I'm Atreya, and I love building DL systems. I've used tensorflow extensively in the past and would love to start contributing so that I can help others have an even better experience! How do I get started with contributions? Thanks!
I also have a few ideas for new implementations that I'd love to get started on if the community is interested in them!
Mustansir Kapasi
Hi everyone, I am Mustansir Kapasi, I am comfortable with C++, Python, Java and have been using Tensorflow 2.0 for developing many of my deep learning projects. I am new to open source if anyone can guide me in the process it would be of great help
Aditya Nikhil
Hey @atreyamaj , I would love to hear your ideas. Like to share them?!
Atreya Majumdar
Hey @AdityaNikhil , so while contributing to existing TF issues, I was thinking of maybe having an implementation of Capsule Networks and Visual Transformers integrated with tf.keras
Do you think this project is feasible?
I'm going through a few more papers to see if we can come up with something else as well!
Kinda similar to importing a model like ResNet50 for example though tf.keras.applications, however it's likely that more input parameters will have to be provided here
Aditya Nikhil
Well I think it's a great idea @atreyamaj. It is feasible, we can integrate with tf models. Let me know how do you plan to do it, I'd love to collaborate.
Atreya Majumdar
Awesome! I'm a bit new to open-source so it would be amazing if you could tell me how the dev process for models normally goes here!
Aditya Nikhil
Hey @atreyamaj, sorry for the late reply. Check this page for research paper code contribution, https://github.com/tensorflow/models/wiki/Research-paper-code-contribution
Aditya Nikhil
In meantime, I'm experimenting tf models my self and trying to fix issues.
Hi everyone! I'm Primesh. I love working in the field of DL and computer vision. I've been using TF extensively for the past 2 years and would like to grow my experience further by becoming a part of the TF community. I have just started to contribute to TF/models.
Shivaditya Meduri
Hi, I am Shiva and I want to contribute to TensorFlow. After making some changes to fix issues in Tensorflow, how do I run the latest tensorflow version from my gitbash while importing tensorflow, When I tried to check the changes in my local computer to see if it fixes the issue, the compiler is always using the stock tensorflow and not incorporating my changes. I would appreciate some tips to solve this issue
Hello everyone,I'm Mudassar I want to contribute to tensorflow.Can anyone guide which are the best issues to solve for the begineer's,so that I can contribute to tensorflow
kartikey rawat
Hello everyone, I am Kartikey Rawat and I am Open Source and ML Enthusiast and I am really want to contribute in tensorflow.js and even I am also advocating tensorflow.js through my speaker session love to be a part of this community
Mustansir Kapasi
Hello everyone and welcome to this Gitter channel of tensorflow! lets see how our contribution helps in the tensorflow repo.
Rohan Raj
Hello everyone, I had some doubts about Convert Tensorflow DL examples to JAX / Flax / Haiku. Any help would be appreciated. Thank you.
Hello, I am looking forward to contribute for the project tensorflow model. Are there are any mentors I can contact with?
how shall i contact the mentors ?
Harshit Verma
Hello, the community! I'm Harshit Verma. I had participated in GSoC'20, where I worked on a C++ project under SCoReLab, and this summer I'm looking forward to work on this incredible project.
Haris Wang
Hello, I am a college student from Sichuan University in China, and I want to apply for GSOC2021. But my college was in the “Supplement No. 4 to Part 744 - ENTITY LIST” by BIS. Will it affect my application?
Drishti Bhasin
Hello, I am Drishti and I will be applying for the upcoming round of Outreachy. I am a sophomore at IIT Roorkee, India and am passionate about ML while being skilled in matlab and python. Hence I would like to contribute to the project 'Create examples for JAX (a new open-source ML library' which seems interesting and suitable for my skillset. Could you please let me know who the mentor for this project this?
Hello Everyone, I am Srikeshram B, A
College Student and I am an ML and DL enthusiastic looking forward to work in GSoC 2021. I would like to contribute to Tensorflow Model Garden. I am very happy to be part of this great community. Will anybody please tell me about contacting mentors for this project?