Emily's internship blog #192
Open
+104
−0
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
💖 Thanks for opening this pull request! 💖 |
|
@EmilyXinyi Could you maybe add a link to or embed your Vlog here? I think it would be a nice addition to this well written blogpost about your internship experience. |
I embedded the video for now, though I am not sure how good it would look on the blog because it's a portrait (vertical) video. Alternatively I can post it on some social media first and link it here. |
|
|
||
| ### Open Source Developement | ||
|
|
||
| I started my contributions by adapting certain metrics (tweedie, mean absolute percentage error etc.) to be Array API compatible under the guidance of my mentor, Olivier. The Array API standard is a cross-library API for array operations on Python, which is designed to improve interoperability and consistency across different array libraries. This also means that scikit-learn algorithms written in NumPy for CPU can work on other hardwares (GPU) with PyTorch or CuPy, greatly improving performance. As I gained more familiarity with the scikit-learn codebase and Array API, I began working on adapting “larger” functions to be Array API compatible, which means a lot more fundamental, a lot more dependencies, a lot more challenging, and a lot more fun. |
There was a problem hiding this comment.
Suggested change
| I started my contributions by adapting certain metrics (tweedie, mean absolute percentage error etc.) to be Array API compatible under the guidance of my mentor, Olivier. The Array API standard is a cross-library API for array operations on Python, which is designed to improve interoperability and consistency across different array libraries. This also means that scikit-learn algorithms written in NumPy for CPU can work on other hardwares (GPU) with PyTorch or CuPy, greatly improving performance. As I gained more familiarity with the scikit-learn codebase and Array API, I began working on adapting “larger” functions to be Array API compatible, which means a lot more fundamental, a lot more dependencies, a lot more challenging, and a lot more fun. | |
| I started my contributions by adapting certain metrics (tweedie, mean absolute percentage error etc.) to be Array API compatible under the guidance of my mentor, [Olivier](https://github.com/ogrisel). The Array API standard is a cross-library API for array operations on Python, which is designed to improve interoperability and consistency across different array libraries. This also means that scikit-learn algorithms written in NumPy for CPU can work on other hardwares (GPU) with PyTorch or CuPy, greatly improving performance. As I gained more familiarity with the scikit-learn codebase and Array API, I began working on adapting “larger” functions to be Array API compatible, which means a lot more fundamental, a lot more dependencies, a lot more challenging, and a lot more fun. |
|
|
||
| ### Chinese Community Outreach | ||
|
|
||
| China has the second largest user group of scikit-learn. As a community, we believe that we can be more inclusive to ease Chinese contribution and do what is necessary to recruit more Chinese contributors. Therefore, I need to find out who and where scikit-learn is being used, if there are other platforms (outside of GitHub) that development is happening, because GitHub tends to be very slow in China, and establish scikit-learn’s official presence in the Chinese community. |
There was a problem hiding this comment.
Suggested change
| China has the second largest user group of scikit-learn. As a community, we believe that we can be more inclusive to ease Chinese contribution and do what is necessary to recruit more Chinese contributors. Therefore, I need to find out who and where scikit-learn is being used, if there are other platforms (outside of GitHub) that development is happening, because GitHub tends to be very slow in China, and establish scikit-learn’s official presence in the Chinese community. | |
| China has the second largest user group of scikit-learn according to documentation web analytics. As a community, we believe that we can be more inclusive to ease Chinese contribution and do what is necessary to onboard more Chinese contributors. Therefore, I need to find out who and where scikit-learn is being used, if there are other platforms (outside of GitHub) that development is happening, because GitHub tends to be very slow in China, and establish scikit-learn’s official presence in the Chinese community. |
|
|
||
| I also had weekly Peer Programming sessions with Loïc and Stefanie, where my piled-up questions from the week outside of Array API would be answered, and I would almost always learn something new about developer tools or programming fundamentals. | ||
|
|
||
| On the Chinese community outreach side, it has always been with the scikit-learn communications team. Here I must give a special shoutout to manager François, who is also part of the communications team, for always being supportive and believing in my outreach efforts, especially because I was nervous doing this kind of task and using Chinese in a professional context for the first time. I also got to interact with [Charlie](https://charlie-xiao.github.io/) (yes, the core-dev Charlie), who is located in China and helped me tremendously with tasks that require physical presence. |
There was a problem hiding this comment.
Suggested change
| On the Chinese community outreach side, it has always been with the scikit-learn communications team. Here I must give a special shoutout to manager François, who is also part of the communications team, for always being supportive and believing in my outreach efforts, especially because I was nervous doing this kind of task and using Chinese in a professional context for the first time. I also got to interact with [Charlie](https://charlie-xiao.github.io/) (yes, the core-dev Charlie), who is located in China and helped me tremendously with tasks that require physical presence. | |
| On the Chinese community outreach side, it has always been with the scikit-learn communications team. Here I must give a special shoutout to manager [François](https://www.linkedin.com/in/françois-goupil/), who is also part of the communications team, for always being supportive and believing in my outreach efforts, especially because I was nervous doing this kind of task and using Chinese in a professional context for the first time. I also got to interact with [Charlie](https://charlie-xiao.github.io/) (yes, the core-dev Charlie), who is located in China and helped me tremendously with tasks that require physical presence. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Blog post about Emily's summer internship
cc: @francoisgoupil