d2l-ai/d2l-tr

I would like to contribute to translate the book into Turkish (if you still need help:)

buzem opened this issue · 4 comments

buzem commented

Merhaba,

If you have still things to do for translating book to Turkish, I would like to contribute this open-source book.I would be glad if you point out the remaining tasks and improvements.Thanks in advance :)

semercim commented

Merhaba Buzem,

Since it is an open-source any contribution are welcome! :) I started the translation around 2 months ago but, as it would be expected, it is so slow. We should organize to avoid unnecessary double-work.

  • My first concern would be to have a consistent language (mostly I mean terminology) throughout the book. First I would suggest whenever you are not certain about a new term, please, kindly refer to the resources I gave in the Translator's note file.

  • Secondly, the current translation is based on Release v0.14.0 of original English text (d2l-ai/d2l-en). Please, kindly, clone or download this specific version (not the latest, unreleased one) into your local machine and use it as the reference in the translation.

  • Please fork d2l-ai/d2l-tr into your github account and start your contributions on your account. I suggest working on chapters as the translation unit. This would be clearer and more tractable. My convention is to create a branch for each individual chapter, then to create a pull request for the chapter to merge it master when it is finished. I include both the original files (with an additional _en extension in the file name) and the translated ones to make it easier for someone who needs to double-check the translation.

  • Of course, you are welcome to correct any misinterpretations or any grammatical and syntax mistakes in the already translated chapters. Having multiple eyes reading the text is the best way to overcome author blindness.

  • I would be glad if you could start with Appendix: Tools for Deep Learning chapter in order to provide the basic information for the people who want to contribute to the book or just want to learn how to use the tools in the book. This would be the next chapter I would plan working on.

  • At the moment, I am translating Linear Neural Networks (Chapter 3) unit. Personally, due to both my academic (had a PhD on machine learning) and professional interests I am particularly interested in Recurrent Neural Networks (Chapter 8), Modern Recurrent Neural Networks (Chapter 9), Attention Mechanisms (Chapter 10) and GANs (Chapter 17). I would be glad if I could be responsible on translating these chapters.

I hope these suggestions and requests also make sense for you.

I am joyfully waiting for your contributions. Thanks for your help offer.

Merhaba,

If you have still things to do for translating book to Turkish, I would like to contribute this open-source book.I would be glad if you point out the remaining tasks and improvements.Thanks in advance :)

buzem commented

Thanks for the answer,

I can only find this version ( 0.14.3) on the following link: https://d2l.ai/d2l-en.pdf. Could you send the 0.14.0 version link to me? Additionally, I am curious about why .md is prefered over .ipynb in both English and Turkish github repos?

semercim commented

Let me try to answer your questions as clear as possible. @barisyasin could provide more information in case I miss something.

d2l-ai is managed like a software project. When it reaches a maturity level or adds new features (such as new chapters) the current state is frozen and an official release is created. You can find all the official releases under:
https://github.com/d2l-ai/d2l-en/releases

This translation uses the latest major official releases in this page, which is Release v0.14.0. From that page, you can download all the materials for all the releases.

There is another way, which I personally prefer. You can clone https://github.com/d2l-ai/d2l-en locally then rewind it to the latest commit used to create the release. This is a little more complex and it just provides you the source files (which are mostly .md files).

Here comes to the answer of your second question. The .md files are the source files for this software project. You can think them as the libraries in your software projects. The same files are used to create the pdf, the website (https://d2l.ai/) and the notebooks. It is a decision given by the initiators of the book. The .md files are used with an automation tool (Jenkins) to automate the creation of these products without any manual intervention.

I hope I could answer your questions.

buzem commented

It became very clear,Thanks for the answer.