Skip to content
Avatar
🤔
🤔

Achievements

Achievements

Highlights

  • Pro
Block or Report

Block or report mariomeissner

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mariomeissner/README.md

Hello!

I'm a Machine Learning researcher and engineer, currently working on Natural Language Processing as a master's student at the University of Tokyo.

Specifically, I'm exploring the shortcut-learning behavior of transformer models, and finding better debiasing techniques to help models avoid spurious correlations in the data. Feel free to visit my blog or have a look at some of my repositories!

If you're struggling to keep your huggingface experiment code organized, try using my opinionated lightning-hydra template! I adapted it from the popular template by ashleve. The most important change is the added support for HuggingFace transformers. Life made easy! Find it below.

Pinned

  1. Forked from ashleve/lightning-hydra-template

    My take on how you should organize your transformer experiments.

    Python 4

  2. AmbiNLI Public

    This is the code for the paper "Embracing Ambiguity: Shifting the Training Target of NLI Models".

    Jupyter Notebook 3

  3. My personal blog!

    Jupyter Notebook

  4. A convolutional neural network similar to U-Net (https://arxiv.org/abs/1505.04597) for medical image segmentation.

    Jupyter Notebook 2

  5. Parsing your Moodle logs, and leaving them ready for data mining!

    Jupyter Notebook

  6. A deep neural network for performing OCR on handwritten Turing Machine definitions.

    Jupyter Notebook 1

196 contributions in the last year

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Mon Wed Fri
Activity overview

Contribution activity

May 2022

Started 1 discussion in 1 repository
3 contributions in private repositories May 7 – May 15

Seeing something unexpected? Take a look at the GitHub profile guide.