{"id":13767616,"url":"https://github.com/proycon/deepfrog","last_synced_at":"2025-04-23T03:52:46.753Z","repository":{"id":37178047,"uuid":"240007492","full_name":"proycon/deepfrog","owner":"proycon","description":"An NLP-suite powered by deep learning","archived":false,"fork":false,"pushed_at":"2023-03-24T23:53:13.000Z","size":2991,"stargazers_count":19,"open_issues_count":18,"forks_count":1,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-04-22T10:21:11.063Z","etag":null,"topics":["deep-learning","deep-neural-networks","dutch","folia","frog","nlp","transformers"],"latest_commit_sha":null,"homepage":"","language":"Rust","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/proycon.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null}},"created_at":"2020-02-12T12:28:13.000Z","updated_at":"2024-05-21T11:44:28.000Z","dependencies_parsed_at":"2024-01-12T00:27:20.912Z","dependency_job_id":"955072ab-9ba3-4e4a-9f17-0ec1adbd22bb","html_url":"https://github.com/proycon/deepfrog","commit_stats":{"total_commits":120,"total_committers":1,"mean_commits":120.0,"dds":0.0,"last_synced_commit":"13d5a594e631230e4f0fc387baff128f8d9024eb"},"previous_names":[],"tags_count":2,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/proycon%2Fdeepfrog","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/proycon%2Fdeepfrog/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/proycon%2Fdeepfrog/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/proycon%2Fdeepfrog/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/proycon","download_url":"https://codeload.github.com/proycon/deepfrog/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":250366688,"owners_count":21418768,"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","dutch","folia","frog","nlp","transformers"],"created_at":"2024-08-03T16:01:10.296Z","updated_at":"2025-04-23T03:52:46.735Z","avatar_url":"https://github.com/proycon.png","language":"Rust","funding_links":[],"categories":["Uncategorized","NLP"],"sub_categories":["Uncategorized"],"readme":"\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"https://raw.githubusercontent.com/proycon/deepfrog/master/logo.png\" width=\"200\" /\u003e\n\u003c/div\u003e\n\n# DeepFrog - NLP Suite\n\n[![Language Machines Badge](http://applejack.science.ru.nl/lamabadge.php/deepfrog)](http://applejack.science.ru.nl/languagemachines/)\n[![Project Status: Suspended -  Initial development has started, but there has not yet been a stable, usable release; work has been stopped for the time being but the author(s) intend on resuming work.](https://www.repostatus.org/badges/latest/suspended.svg)](https://www.repostatus.org/#suspended)\n\n\n## Introduction\n\nDeepFrog aims to be a (partial) successor of the Dutch-NLP suite [Frog](https://languagemachines.github.io/frog).\nWhereas the various NLP modules in Frog wre built on k-NN classifiers, DeepFrog builds on deep learning techniques and\ncan use a variety of neural transformers.\n\nOur deliverables are multi-faceted:\n\n1. Fine-tuned neural network **models** for Dutch NLP that can be compared with [Frog](https://languagemachines.github.io/frog) and are\n    directly usable with [Huggingface's Transformers](https://github.com/huggingface/transformers) library for Python (or\n    [rust-bert](https://github.com/guillaume-be/rust-bert) for Rust).\n2. **Training pipelines** for the above models (see ``training``).\n3. A **software tool** that integrates multiple models (not just limited to dutch!) and provides a single pipeline solution for end-users.\n    * with full support for [FoLiA XML](https://proycon.github.io/folia) input/output.\n    * usage is not limited to the models we provide\n\n## Installation\n\nDeepFrog and all its dependencies are included as an extra in [LaMachine](https://proycon.github.io/LaMachine), which is the easiest\nway to install it. Within lamachine, do:\n\n  ``lamachine-add deepfrog \u0026\u0026 lamachine-update``\n\nOtherwise, simply install DeepFrog using Rust's package manager:\n\n```\ncargo install deepfrog\n```\n\nNo cargo/rust on your system yet? Do ``sudo apt install cargo`` on Debian/ubuntu based systems, ``brew install rust`` on mac, or use [rustup](https://rustup.rs/).\n\nIn order to run the DeepFrog command-line-tool, you need to have the C++ library [libtorch](https://pytorch.org/) installed.\nDownload it from https://pytorch.org/ ; make sure you select *package: libtorch* there! You don't need the rest of\npytorch.\n\n\nTo use our models directly with the [Huggingface's Transformers](https://github.com/huggingface/transformers) library\nfor Python, you merely need that library, models will be automatically downloaded and installed as you invoke them. The\nDeepFrog command-line-tool is not used in this workflow.\n\n\n## Models\n\nWe aim to make available various models for Dutch NLP.\n\n#### RobBERT v1 Part-of-Speech (CGN tagset) for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/robbert-pos-cased-deepfrog-nld\n\nUses pre-trained model [RobBERT](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/) (a Roberta model), fine-tuned on\npart-of-speech tags with the full corpus as also used by Frog. Uses the tag set of Corpus Gesproken Nederlands (CGN), this\ncorpus constitutes a subset of the training data.\n\nTest Evaluation:\n\n```\nf1 = 0.9708171206225681\nloss = 0.07882563415198372\nprecision = 0.9708171206225681\nrecall = 0.9708171206225681\n```\n\n#### RobBERT v2 Part-of-Speech (CGN tagset) for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/robbert2-pos-cased-deepfrog-nld\n\nUses pre-trained model [RobBERT v2](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/) (a Roberta model), fine-tuned on\npart-of-speech tags with the full corpus as also used by Frog. Uses the tag set of Corpus Gesproken Nederlands (CGN), this\ncorpus constitutes a subset of the training data.\n\n```\nf1 = 0.9664560038891591\nloss = 0.09085878504153627\nprecision = 0.9659863945578231\nrecall = 0.9669260700389105\n```\n\n#### BERT Part-of-Speech (CGN tagset) for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/bert-pos-cased-deepfrog-nld\n\nUses pre-trained model [BERTje](https://github.com/wietsedv/bertje) (a BERT model), fine-tuned on\npart-of-speech tags with the full corpus as also used by Frog. Uses the tag set of Corpus Gesproken Nederlands (CGN), this\ncorpus constitutes a subset of the training data.\n\nTest Evaluation:\n\n```\nf1 = 0.9737354085603113\nloss = 0.0647074995296342\nprecision = 0.9737354085603113\nrecall = 0.9737354085603113\n```\n\n#### RobBERT SoNaR1 Named Entities for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/robbert-ner-cased-sonar1-nld\n\nUses pre-trained model [RobBERT](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/) (a Roberta model), fine-tuned on\nNamed Entities from the SoNaR1 corpus (as also used by Frog). Provides basic PER,LOC,ORG,PRO,EVE,MISC tags.\n\n\nTest Evaluation (**note:** *this is a simple token-based evaluation rather than entity based!*)\n\n```\nf1 = 0.9170731707317074\nloss = 0.023864904676364467\nprecision = 0.9306930693069307\nrecall = 0.9038461538461539\n```\n\n**Note:** the tokenisation in this model is English rather than Dutch\n\n#### RobBERT v2 SoNaR1 Named Entities for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/robbert2-ner-cased-sonar1-nld\n\nUses pre-trained model [RobBERT (v2)](https://people.cs.kuleuven.be/~pieter.delobelle/robbert/) (a Roberta model), fine-tuned on\nNamed Entities from the SoNaR1 corpus (as also used by Frog). Provides basic PER,LOC,ORG,PRO,EVE,MISC tags.\n\n```\nf1 = 0.8878048780487806\nloss = 0.03555946223787032\nprecision = 0.900990099009901\nrecall = 0.875\n```\n\n\n#### BERT SoNaR1 Named Entities for Dutch\n\nModel page with instructions: https://huggingface.co/proycon/bert-ner-cased-sonar1-nld\n\nUses pre-trained model [BERTje](https://github.com/wietsedv/bertje) (a BERT model), fine-tuned on\nNamed Entities from the SoNaR1 corpus (as also used by Frog). Provides basic PER,LOC,ORG,PRO,EVE,MISC tags.\n\nTest Evaluation (**note:** *this is a simple token-based evaluation rather than entity based!*)\n\n```\nf1 = 0.9519230769230769\nloss = 0.02323892477299803\nprecision = 0.9519230769230769\nrecall = 0.9519230769230769\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fproycon%2Fdeepfrog","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fproycon%2Fdeepfrog","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fproycon%2Fdeepfrog/lists"}