{"id":15539809,"url":"https://github.com/salamanderxing/mate","last_synced_at":"2025-11-05T01:02:26.644Z","repository":{"id":64868684,"uuid":"577278840","full_name":"SalamanderXing/mate","owner":"SalamanderXing","description":"🧉 the ultimate deep learning project management and sharing tool","archived":false,"fork":false,"pushed_at":"2023-10-15T10:05:59.000Z","size":76562,"stargazers_count":4,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-04-19T03:14:24.700Z","etag":null,"topics":["deep-learning","deep-neural-networks","deeplearning","jax","machine-learning","project-template","pytorch","pytorch-lightning","reproducibility","research","tensorflow"],"latest_commit_sha":null,"homepage":"","language":"HTML","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/SalamanderXing.png","metadata":{"files":{"readme":"README.md","changelog":"changes.txt","contributing":null,"funding":null,"license":"LICENSE.txt","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2022-12-12T11:41:06.000Z","updated_at":"2023-02-12T14:43:08.000Z","dependencies_parsed_at":"2023-10-16T03:03:09.402Z","dependency_job_id":null,"html_url":"https://github.com/SalamanderXing/mate","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SalamanderXing%2Fmate","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SalamanderXing%2Fmate/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SalamanderXing%2Fmate/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/SalamanderXing%2Fmate/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/SalamanderXing","download_url":"https://codeload.github.com/SalamanderXing/mate/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":250468279,"owners_count":21435454,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deep-learning","deep-neural-networks","deeplearning","jax","machine-learning","project-template","pytorch","pytorch-lightning","reproducibility","research","tensorflow"],"created_at":"2024-10-02T12:11:33.307Z","updated_at":"2025-11-05T01:02:26.631Z","avatar_url":"https://github.com/SalamanderXing.png","language":"HTML","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Maté 🧉\n\nMate is a tool designed to improve reproducibility and facilitate development in\ndeep learning. It is a command line tool that offers a variety of features to\nhelp you manage your project, including\n\n- Validating the structure of your project\n- Visualizing/summarizing your project\n- Remote execution of experiments\n- Summarizing your results\n- Running/testing your experiments.\n- Managment of Python virtual environment\n\n\u003cdiv style=\"background-color: #ffffcc; border-left: 6px solid #ffeb3b; padding: 0.5em;\"\u003e\n  ⚠️  Mate is currently under heavy development. Contact me for more details.\n\u003c/div\u003e\n\n\u003c!----\u003e\n\u003c!-- It also creates a universal template for deep learning projects. --\u003e\n\u003c!-- In addition, any project developed with Mate on a public repository gets automatically listed on MateHub. This website is a browser for finding and reusing components created by others (or yourself). --\u003e\n\u003c!-- Mate is compatible with any python deep learning framework, such as PyTorch, JAX, and TensorFlow/Keras, since it leverages Python features. --\u003e\n\u003c!----\u003e\n\u003c!----\u003e\n\u003c!-- ## [documentation](https://salamanderxing.github.io/mate) --\u003e\n\u003c!----\u003e\n\u003c!-- Mate is developed in collaboration with the University of Amsterdam. --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## Installation 🔌 --\u003e\n\u003c!----\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_8b3d5a8def640d1dc9b67d83aff7397e.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## Example Projects --\u003e\n\u003c!----\u003e\n\u003c!----\u003e\n\u003c!-- ### PyTorch Lightning --\u003e\n\u003c!----\u003e\n\u003c!-- - [MNIST Classifier](https://github.com/SalamanderXing/pytorch-lightning-mnist) --\u003e\n\u003c!----\u003e\n\u003c!-- ### JAX (with Flax) --\u003e\n\u003c!-- [JAX](https://github.com/google/jax) is the (relatively) new framework by Google. That uses just-in-time compilation to improve performance of your neural network. --\u003e\n\u003c!-- These projects are based on this [amazing repo](https://github.com/phlippe/uvadlc_notebooks/tree/master/docs/tutorial_notebooks/JAX). --\u003e\n\u003c!----\u003e\n\u003c!-- - [CIFAR10 Autoecoder](https://github.com/SalamanderXing/jax-ae) --\u003e\n\u003c!-- - [Graph Neural Networks](https://github.com/SalamanderXing/jax-gnn) --\u003e\n\u003c!-- - [Normalizing Flow](https://github.com/SalamanderXing/jax-normalizing-flow) --\u003e\n\u003c!-- - [Inception, ResNet, DenseNet](https://github.com/SalamanderXing/jax-inception-resnet-densenet) --\u003e\n\u003c!-- - [Autoregressive Image Modeling](https://github.com/SalamanderXing/jax-autoregressive-image-modeling) --\u003e\n\u003c!-- - [Transformers for text classification](https://github.com/SalamanderXing/jax-transformers) --\u003e\n\u003c!-- - [Transformers for anomaly detection](https://github.com/SalamanderXing/jax-anomaly-detection) --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## Quick Start ⚡ --\u003e\n\u003c!-- This example will walk you through training and showing results of your model on a Pytorch-Lightning example. --\u003e\n\u003c!----\u003e\n\u003c!-- First let's clone the project, for example:  --\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_49449704768719c08e05230ff2ab1f5b.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!-- Then go to the project directory: --\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_921a8028dd628088e0c41e17f4ab2d06.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!-- Then run: --\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/None_cacf09445830e5e547952f44e09ae2a6.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!-- This will give you an overview of your project and its components. Besides, it will create a local python virtual environment and install dependencies of this project. It also tells you where your components have issues. --\u003e\n\u003c!----\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/exec_5ac24db831400cf68943454e2be32f48.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!-- then we can train our experiment: --\u003e\n\u003c!----\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_27e5978a0d23c95eaa27ace2684499f6.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!-- You should see all the training logs. --\u003e\n\u003c!-- If you now do again: --\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_cacf09445830e5e547952f44e09ae2a6.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!-- You should notice a 💪 next to the experiment. That means that the training was successful: --\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/exec_5ac24db831400cf68943454e2be32f48.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!----\u003e\n\u003c!-- Finally, to visualize our results: --\u003e\n\u003c!----\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/bash_eda9dcc34a8ecfc1ae5dc9aafab9c28d.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!-- \u003cp align=\"center\" style=\"\"\u003e --\u003e\n\u003c!--     \u003cimg src=\"./imgs/exec_e2b436d13e17ff69e83786cac9a87b76.svg\" style=\"max-width:550px\" alt=\"Your Image\"\u003e --\u003e\n\u003c!-- \u003c/p\u003e --\u003e\n\u003c!----\u003e\n\u003c!-- Mate has inferred the dataset and will group our experiments according to that and put them in the same table. --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## [MateHub](https://salamanderxing.github.io/matehub/) --\u003e\n\u003c!----\u003e\n\u003c!-- Before creating a new module (trainer, data_loader, model), you might want to head over to this site and see if there is anything that fits your need or allows you to not start from scratch. --\u003e\n\u003c!----\u003e\n\u003c!-- ### How it works --\u003e\n\u003c!----\u003e\n\u003c!-- All mate projects on public GitHub repo (published by anyone) will be automatically listed on MateHub. It works by using the GitHub rest API. --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## Comparison to familiar tools --\u003e\n\u003c!----\u003e\n\u003c!-- - *[Weights \u0026 Biases](https://wandb.ai/site)*, *[Tensorboard](https://www.tensorflow.org/tensorboard)*  is a logger and allows model weights sharing as well. Mate does not attempt to replace logger's functionalities. Use the logger your like best :) That would happend probably inside your trainer module.  --\u003e\n\u003c!-- - *[Monai](https://github.com/Project-MONAI/MONAI)*: Focuses on medical imaging and provides pretrained models as well as preprocessing pipelines --\u003e\n\u003c!-- - *[Ivy](https://github.com/unifyai/ivy)*: Provides a unified tensor type that work with all backends (frameworks). Works with mate! --\u003e\n\u003c!-- - *[THINGSvision](https://github.com/ViCCo-Group/thingsvision)*: Provides a set pretrained models for analysis of their activation. In particular to compare them with brain activations. --\u003e\n\u003c!-- - *[HuggingFace](https://huggingface.co/), [Model Zoo](https://modelzoo.co/)*: These frameworks focus on sharing pretrained models. Mate instead focuses on (among other things) sharing the model components. --\u003e\n\u003c!----\u003e\n\u003c!-- --- --\u003e\n\u003c!----\u003e\n\u003c!-- ## Contact 🤝  --\u003e\n\u003c!----\u003e\n\n## Contacts\n\n- [Email](mailto:g.zani@uva.nl)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsalamanderxing%2Fmate","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsalamanderxing%2Fmate","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsalamanderxing%2Fmate/lists"}