AdaptNLP

The official AdaptNLP twitter account!
Github: @TheZachMueller

[Zach Mueller @TheZachMueller]

It's a very big day for me today, as I'm officially releasing version 0.2.3 for the @AdaptNLP project. With it comes a slew of changes, but what exactly?

A preview:
#nbdev, @fastdotai, #fastcore, and something called a ModelHub for both #FlairNLP and @huggingface, let's dive in:

[Andrew Chang @achang1618]

@AdaptNLP 0.2.1 out with additional fine-tuning and training tutorial and notebooks here https://t.co/tujLHTIznh

Check out how to use the phenomenal @huggingface datasets and transformers libraries for streamlined fine-tuning and training of NLP models

#NLProc #Python #ML #AI

[Andrew Chang @achang1618]

@AdaptNLP
v0.2.0 is pushed and out! Check out the release notes here: https://t.co/H94ecGLz3g

New features:
- New training framework integrated with transformers Trainer
- @huggingface nlp package is now a MAJOR dependency
-Docs: https://t.co/11B4Y7jtcQ

#NLProc #NLP #python

[Andrew Chang @achang1618]

@omarsar0 had an awesome virtual tutorial at ODSC the other day that provided a great 3 part course walkthrough at https://t.co/qiuKxmQQ7c!

Great NLP start to the conference and super excited to present our @AdaptNLP workshop on Friday with @bsacash! #ODSCVirtual

[Novetta - a part of Accenture Federal Services @Novettasol]

ODSC East is Virtual! And we're still going! Catch Novettans Andrew, Brian, and Shauna - training and explaining. From their homes to yours. Read our latest blog to prep for our AdaptNLP workshop: https://t.co/A6xXjHnsTf #NLP @odsc @AdaptNLP @achang1618 @bsacash

[Andrew Chang @achang1618]

Just released AdaptNLP v0.1.4!
You can now run its new summarization and translation modules with the T5 and Bart language models from @huggingface's latest release 🤗

Tutorial guides, colab notebooks, and API documentation are available at https://t.co/0VDI8Fls9b!
#NLProc

[Novetta - a part of Accenture Federal Services @Novettasol]

ELECTRA, Stanza, and SLIDE may bring back memories from the 90's, but the technology behind these names is the latest and greatest in Machine Learning. Learn more in this month's issue of Baseline: https://t.co/PZBMfBoriz @AdaptNLP @Spacenet_AI @GoogleAI @TensorFlow #NLP

[Andrew Chang @achang1618]

The "Auto" modules in @huggingface Transformers have been a huge part of what made model, configs, and tokenizer loading so seamless. Check out their new v2.6.0 release that now includes the BART model!

Here's a snippet of how @AdaptNLP utilized the Auto Classes:

[Shivanand @snrspeaks]

🔥Easily implement BERT, RoBERT, XLNet and other transformer models for your next NLP tasks with these libraries:
⚡AdaptNLP
⚡Simple Transformers
⚡Ktrain
⚡Fast Bert
⚡Flair

#DataScience
#MachineLearning
#DeepLearning
#nlp
#ArtificialIntelligence

https://t.co/SxrRnXOzqn

[ODSC (Open Data Science Conference) AI @_odsc]

Let's explore what it would take to streamline the implementation of state-of-the-art models for different NLP tasks allowing for quick use by practitioners. #DataScience #NLP #LanguageProcessing https://t.co/f5ezVvilub

[AdaptNLP @AdaptNLP]

You can now easily try AdaptNLP on Kaggle in the Real or Not? NLP with Disaster Tweets competition. Get started with this kernel: https://t.co/kJoGlFX6B5

[ODSC (Open Data Science Conference) AI @_odsc]

Let's explore what it would take to streamline the implementation of state-of-the-art models for different NLP tasks allowing for quick use by practitioners. #DataScience #NLP #LanguageProcessing https://t.co/byGvB7uGzT

[AdaptNLP @AdaptNLP]

New updates for release v0.1.3 at https://t.co/ngpWcR3QfN, feel free to check it out!

[Andrew Chang @achang1618]

AdaptNLP v0.1.3 is released!

@huggingface's awesome public models are now available to run using @AdaptNLP's sequence classification and QA modules with mini batch inference

Check out the docs for deploying these models with FastAPI and GPUs! https://t.co/PDAMbYTbPw

#NLProc

[Andrew Chang @achang1618]

Just pinned FastAPI's v0.49 in @AdaptNLP. Looking forward to integrating some of FastAPI's latest updates from @tiangolo in our roadmap for serving NLP models! #NLProc #fastapi

[AdaptNLP @AdaptNLP]

Come visit us at @odsc East and see how you can apply the latest NLP capabilities to your projects with the AdaptNLP library!

#opensource #odsc #python #PyTorch #NLProc

[Novetta - a part of Accenture Federal Services @Novettasol]

Catch @achang1618 and @bsacash's talk: “State-of-the-art NLP Made Easy" at #ODSCeast, April 13-17 in Boston. In the meantime, check out Novetta's first open source framework, @AdaptNLP, on GitHub. https://t.co/fVbbRrClhX

[Sebastian Ruder @seb_ruder]

There have been many cool datasets recently, which will hopefully spur QA research in other languages:
- 🇨🇳DuReader: https://t.co/8yNoxg91MA DRCD: https://t.co/lH2Nk0DIET
- 🇰🇷KorQuAD: https://t.co/RcLEoSi1oK
- 🇷🇺SberQuAD: https://t.co/fHoEhBgz4l
- 🇫🇷FQuAD: https://t.co/NdsdxAKm72

[Andrew Chang @achang1618]

These slides are great! @tiangolo's FastAPI has provided an amazing tool for serving ML models in python. @AdaptNlp uses FastAPI to serve pre-trained NLP models like NER and Question Answering as well as custom trained models. Try it out here! https://t.co/u1qr11tju2