Curtis G. Northcutt

benchmarking-keras-pytorch. 🔥 Reproducibly benchmarking Keras and PyTorch models

hypopt. ⏸ Parallelized hyper-param optimization with validation set, not crossval

rankpruning. 🧹 Formerly for binary classification with noisy labels. Replaced by cleanlab.

cleanlab. Official cleanlab repo is at https://github.com/cleanlab/cleanlab

confidentlearning-reproduce. Official data release to reproduce Confident Learning paper results

cnn-gpu-benchmarks. Latest (2020) CNN and GPU Benchmarks on ImageNet and CIFAR

ieee-keywords. IEEE Computer Society Keywords to Organize Knowledge

veneer. 😎 Obscures commit history on GitHub.

reliablity_framework_for_rag. Demo showing how the Trustworthy Language Model add reliability to LLM outputs and improves RAG, agents, and data enrichment worfklows. can be used to improve fine-tuning of LLMs, accuracy of LLM outputs, and smart routing for RAG and agents.

forum-diversification. Add diversity to the order of comments in forums!

systemspecs. 💻 Prints GPU, CPU, RAM, storage, etc. for linux/unix systems.

cgnorthcutt.github.io. Curtis G. Northcutt's personal website.

ScrapeGitHubEmails. Python

Awesome-Learning-with-Label-Noise. A curated list of resources for Learning with Noisy Labels

coteaching_plus. ICML'19: How does Disagreement Help Generalization against Label Corruption?

SemanticTextDB. OpenEdge ABL

documentation-theme-jekyll. A Jekyll-based theme designed for documentation and help systems. See the link for detailed instructions on setting up and configuring everything.

streamlit-demo. Uber app demo using streamlit

snorkel. A system for quickly generating training data with weak supervision

ribbit.

l2e-docker. Shell

creative-theme-jekyll. HTML

pomdp. The official website of PomDP the PhD rapper.

iteround. Sum-safe Rounding for Iterables

pytorch-cnn-finetune. Fine-tune pretrained Convolutional Neural Networks with PyTorch

scripts. misc useful things

python-cpu-stress-test-benchmark. Benchmark your CPU multi-thread and single-thread speed without needing sudo.

imagenet-multiGPU.torch. an imagenet example in torch.

m2r. Markdown to reStructuredText converter

cuml. cuML - RAPIDS Machine Learning Library

dotfiles. Uses dotbot to handle my dotfiles.

SCELoss-Reproduce. Reproduce Results for ICCV2019 "Symmetric Cross Entropy for Robust Learning with Noisy Labels" https://arxiv.org/abs/1908.06112

tox_cov_travis_pytest_framework. Python

mixup-cifar10. mixup: Beyond Empirical Risk Minimization

x. playground for testing stuff on github

gitignore. A collection of useful .gitignore templates

project-pages. Fork this repo for a quick start. If "Project Timeline" or "License" appeared on your nav bar, Look Below!

scholar.py. A parser for Google Scholar, written in Python

Replica-Dataset. The Replica Dataset v1 as published in https://arxiv.org/abs/1906.05797 .

hydejack. "Best Jekyll Theme by a Mile"

templates. Document templates for open-source projects (README, CONTRIBUTING, GitHub templates)

Co-teaching. NeurIPS'18: Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels

PyTorch_CIFAR10. Pretrained TorchVision models on CIFAR10 dataset (with weights)

EgoCom-Dataset. EgoCom: A Multi-person Multi-modal Egocentric Communications Dataset

models. Models and examples built with TensorFlow

sentence-similarity. This repository contains various ways to calculate sentence vector similarity using NLP models

imagenet-simple-labels. Simpler human-readable labels for ImageNet 🏷

quickdraw-dataset. Documentation on how to access and use the Quick, Draw! Dataset.

noisy_label_understanding_utilizing. ICML 2019: Understanding and Utilizing Deep Neural Networks Trained with Noisy Labels

cameo. A Python Pandas library for cheating, collaboration, proficiency, and retention detection in edX courses ingested with Google BigQuery.