All Projects → aquadzn → Gpt2 French

aquadzn / Gpt2 French

Licence: other
GPT-2 French demo | Démo française de GPT-2

Projects that are alternatives of or similar to Gpt2 French

Indonesian Language Models
Indonesian Language Models and its Usage
Stars: ✭ 64 (+36.17%)
Mutual labels:  jupyter-notebook, language-model, transformer
Vietnamese Electra
Electra pre-trained model using Vietnamese corpus
Stars: ✭ 55 (+17.02%)
Mutual labels:  jupyter-notebook, language-model, transformer
Gpt2 Chinese
Chinese version of GPT2 training code, using BERT tokenizer.
Stars: ✭ 4,592 (+9670.21%)
Mutual labels:  text-generation, transformer
Deepsvg
[NeurIPS 2020] Official code for the paper "DeepSVG: A Hierarchical Generative Network for Vector Graphics Animation". Includes a PyTorch library for deep learning with SVG data.
Stars: ✭ 403 (+757.45%)
Mutual labels:  jupyter-notebook, transformer
Pytorch Original Transformer
My implementation of the original transformer model (Vaswani et al.). I've additionally included the playground.py file for visualizing otherwise seemingly hard concepts. Currently included IWSLT pretrained models.
Stars: ✭ 411 (+774.47%)
Mutual labels:  jupyter-notebook, transformer
Azureml Bert
End-to-End recipes for pre-training and fine-tuning BERT using Azure Machine Learning Service
Stars: ✭ 342 (+627.66%)
Mutual labels:  jupyter-notebook, language-model
Question generation
Neural question generation using transformers
Stars: ✭ 356 (+657.45%)
Mutual labels:  jupyter-notebook, transformer
Tsai
Time series Timeseries Deep Learning Pytorch fastai - State-of-the-art Deep Learning with Time Series and Sequences in Pytorch / fastai
Stars: ✭ 407 (+765.96%)
Mutual labels:  jupyter-notebook, transformer
Transformer
Implementation of Transformer model (originally from Attention is All You Need) applied to Time Series.
Stars: ✭ 273 (+480.85%)
Mutual labels:  jupyter-notebook, transformer
Sentiment analysis fine grain
Multi-label Classification with BERT; Fine Grained Sentiment Analysis from AI challenger
Stars: ✭ 546 (+1061.7%)
Mutual labels:  jupyter-notebook, language-model
Nlp Paper
NLP Paper
Stars: ✭ 484 (+929.79%)
Mutual labels:  language-model, transformer
Awesome Bert Nlp
A curated list of NLP resources focused on BERT, attention mechanism, Transformer networks, and transfer learning.
Stars: ✭ 567 (+1106.38%)
Mutual labels:  language-model, transformer
Gpt2client
✍🏻 gpt2-client: Easy-to-use TensorFlow Wrapper for GPT-2 117M, 345M, 774M, and 1.5B Transformer Models 🤖 📝
Stars: ✭ 322 (+585.11%)
Mutual labels:  text-generation, transformer
Dab
Data Augmentation by Backtranslation (DAB) ヽ( •_-)ᕗ
Stars: ✭ 294 (+525.53%)
Mutual labels:  jupyter-notebook, transformer
Bert Multitask Learning
BERT for Multitask Learning
Stars: ✭ 380 (+708.51%)
Mutual labels:  jupyter-notebook, transformer
Demo Chinese Text Binary Classification With Bert
Stars: ✭ 276 (+487.23%)
Mutual labels:  jupyter-notebook, transformer
Neural sp
End-to-end ASR/LM implementation with PyTorch
Stars: ✭ 408 (+768.09%)
Mutual labels:  language-model, transformer
Getting Things Done With Pytorch
Jupyter Notebook tutorials on solving real-world problems with Machine Learning & Deep Learning using PyTorch. Topics: Face detection with Detectron 2, Time Series anomaly detection with LSTM Autoencoders, Object Detection with YOLO v5, Build your first Neural Network, Time Series forecasting for Coronavirus daily cases, Sentiment Analysis with BERT.
Stars: ✭ 738 (+1470.21%)
Mutual labels:  jupyter-notebook, transformer
amrlib
A python library that makes AMR parsing, generation and visualization simple.
Stars: ✭ 107 (+127.66%)
Mutual labels:  text-generation, transformer
text-generation-transformer
text generation based on transformer
Stars: ✭ 36 (-23.4%)
Mutual labels:  text-generation, transformer

Edit Mai 2020

Ce repo est un peu en désordre et mal expliqué sur certains points.

Ayant accès au Tensorflow Research Cloud Program, je compte peut être entraîner from scratch (et non fine-tuner comme il est question ici) sur un gros dataset en français. Si cela abouti, j'archiverai ce qui est ici actuellement et j'écrirai une meilleure doc 👍

GPT-2 🇫🇷

Modèle GPT-2 d'OpenAI entraîné sur quatres différents jeux de données en français:

  • Livres en français
  • Scripts de films français
  • Compte-rendus des débats parlementaires
  • Tweet d'Emmanuel Macron

permettant de faire de la génération de texte.

Une démo utilisant pour chaque jeu de données un modèle GPT-2 Small (124M) est disponible à l'adresse suivante: https://gpt2.williamjacques.fr (service down)

Usage

Cloner ce repo

git clone https://github.com/aquadzn/gpt2-french.git

Tensorflow et gpt-2-simple sont necéssaires afin de fine-tuner GPT-2. Créer un environnement puis installer les deux packages pip install tensorflow==1.14 gpt-2-simple.

Un script et un notebook sont disponibles dans le dossier src afin de fine-tuner GPT-2 sur vos propres datasets. L'output de chaque entraînement, c'est à dire le dossier checkpoint/run1, est à mettre dans gpt2-model/model1 model2 model3 etc...

image

Vous pouvez lancer le script deploy_cloudrun.sh afin de déployer tous vos différents modèles (dans gpt2-model) d'un coup. Il faut cependant avoir déjà initialisé l'outil CLI gcloud (Cloud SDK).

Modèles

Releases

Crédits

Basé sur les repos GPT2-Simple et GPT2-CloudRun de minimaxir

Licence

MIT

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].