360K subscribers
4.33K photos
816 videos
17 files
4.81K links
Погружаемся в машинное обучение и Data Science

Показываем как запускать любые LLm на пальцах.

По всем вопросам - @haarrp

@itchannels_telegram -🔥best channels

Реестр РКН: clck.ru/3Fmqri
Download Telegram
Galaxy Zoo: Classifying Galaxies with Crowdsourcing and Active Learning

In this tutorial you will know how to use crowdsourcing and machine learning to investigate how galaxies evolve by classifying millions of galaxy images.

https://blog.tensorflow.org/2020/05/galaxy-zoo-classifying-galaxies-with-crowdsourcing-and-active-learning.html

Code: https://github.com/mwalmsley/galaxy-zoo-bayesian-cnn/blob/88604a63ef3c1bd27d30ca71e0efefca13bf72cd/zoobot/active_learning/acquisition_utils.py#L81
Graph Structure Learning for Robust Graph Neural Networks

A general framework Pro-GNN, which can jointly learn a structural graph and a robust graph neural network model from the perturbed graph guided by these properties.

Github: https://github.com/ChandlerBang/Pro-GNN

Paper: https://arxiv.org/abs/2005.10203
SymJAX: symbolic CPU/GPU/TPU programming

SymJAX is a symbolic programming version of JAX simplifying graph input/output/updates and providing additional functionalities for general machine learning and deep learning applications.

docs: https://symjax.readthedocs.io/en/latest/

github: https://github.com/RandallBalestriero/SymJAX

pdf: https://arxiv.org/pdf/2005.10635v1.pdf
Evaluating Natural Language Generation with BLEURT

BLEURT (Bilingual Evaluation Understudy with Representations from Transformers) builds upon recent advances in transfer learning to capture widespread linguistic phenomena, such as paraphrasing

https://ai.googleblog.com/2020/05/evaluating-natural-language-generation.html

Github: https://github.com/google-research/bleurt

Paper: https://arxiv.org/abs/2004.04696
DETR: End-to-End Object Detection with Transformers

PyTorch training code and pretrained models for DETR The main ingredients of the new framework, called DEtection TRansformer or DETR, are a set-based global loss that forces unique predictions via bipartite matching, and a transformer encoder-decoder architecture.

Github: https://github.com/facebookresearch/detr

Paper: https://arxiv.org/abs/2005.12872v1

Code: https://colab.research.google.com/github/facebookresearch/detr/blob/colab/notebooks/detr_demo.ipynb
How MTS used smart contract to build a system for selecting best technological projects.

https://habr.com/ru/company/ru_mts/blog/504058
Towards computer-aided severity assessment: training and validation of deep neural networks for geographic extent and opacity extent scoring of chest X-rays for SARS-CoV-2 lung disease severity

The COVID-Net models provided here are intended to be used as reference models that can be built upon and enhanced as new data becomes available.

Github: https://github.com/lindawangg/COVID-Net

Paper: https://arxiv.org/abs/2005.12855v1
👍1
Segmentation Loss Odyssey

Loss functions for image segmentation

Github: https://github.com/JunMa11/SegLoss

Paper: https://arxiv.org/abs/2005.13449v1
NeuralPy

NeuralPy: A Keras like deep learning library works on top of PyTorch PyTorch is an open-source machine learning framework that accelerates the path from research prototyping to production deployment developed by Facebook runs on both CPU and GPU.

Github: https://github.com/imdeepmind/NeuralPy

Project: https://neuralpy.imdeepmind.com/
Text Mining in Python: Steps and Examples

This blog summarizes text preprocessing and covers the NLTK steps including Tokenization, Stemming, Lemmatization, POS tagging, Named entity recognition and Chunking.

https://www.kdnuggets.com/2020/05/text-mining-python-steps-examples.html
Acme: A research framework for reinforcement learning

Acme strives to expose simple, efficient, and readable agents, that serve both as reference implementations of popular algorithms and as strong baselines, while still providing enough flexibility to do novel research

Github: https://github.com/deepmind/acme

Paper: https://arxiv.org/abs/2006.00979
A Smooth Representation of SO(3) for Deep Rotation Learning with Uncertainty

In this work presented a novel symmetric matrix representation of rotations that is singularity-free and requires marginal computational overhead

Website: https://papers.starslab.ca/bingham-rotation-learning/

Paper: https://arxiv.org/abs/2006.01031

Github: https://github.com/utiasSTARS/bingham-rotation-learn