renatoviolin / next_word_prediction Goto Github PK

Using transformers to predict next word and predict <mask> word

Python 40.35% CSS 4.34% JavaScript 17.02% HTML 38.29%

next_word_prediction's Introduction

Next word prediction

Simple application using transformers models to predict next word or a masked word in a sentence.

The purpose is to demo and compare the main models available up to date.

The first load take a long time since the application will download all the models. Beside 6 models running, inference time is acceptable even in CPU.

Application

This app implements two variants of the same task (predict token). The first one consider the is at end of the sentence, simulating a prediction of the next word of the sentece.

The second variant is necessary to include a token where you want the model to predict the word.

Running

cd web-app
python app.py

Open your browser http://localhost:8000

next_word_prediction's People

Contributors

Stargazers

Watchers

Forkers

wkryst abhinavm24 ashishpatel26 amirstudy rprilepskiy lozanocelia manikant92 gyanachand1 kiminh yiqiao-yin ravisrirangam yaoxinbin generalzh indyer25 stefanushinardi iflybird keanukf zeus1503 amanjakhetiya ejhortala gopalchettri abhijitdalavi puzuwe dineshdevpandey uunal krishnakarthi deepchatterjeevns tsivaguru mnpk 210010 raman-verma hsali qqq-tech trendingtechnology sehsantaher yiyiwang515 gazzola amit6604 rojinva praveenjoshi01 suitup autosoft-dev em-biet-code ruhulali dendisuhubdy vishalmishra14 hbcbh1999 kaburelabs antoniuschristiyanto nomadtomas vidush5 avsolatorio marcelomata dataorz kc2fresh sudoknight inspiringpeople rainmaker29 krishna4291083 ferhimedamine srinivasanbigdata hadryan lyuji282 hxu927 barryzm varishtsg hhy5277 dsp6414 shubhra2503 tchigher abhisheksinha1506 zhiephieforks cy-dev-tex stephen-kamau fighting41love rafmacalaba saurabhgupta17888 hospice imozart pandeynandancse ak799 roholazandie kenakpatel ekanshbhatnagar raphael-sch kumblemadhu rogervaas rahulchetlangia idannik jasondeving qiming-zou schoudhary101 arryboom armheb aakashchauhan90 sra1github rexhaif gitlost-murali a-chelimo godpgf

next_word_prediction's Issues

Google Colab Version. Flask issue 📍

Very grateful for the interface. Thank you.
I'm trying to make this interface more accessible to others by providing a Colab version of it. Unfortunately, I'm having a bit of an issue with getting the flask server to be integrated in Colab. This is as far as I've gotten. Maybe you can take it from there?

Colab of Next Word Prediction

Current issue is that "ngrok" is causing the error:
TypeError: new_run() got an unexpected keyword argument 'host'
I'm not familiar enough with flask or ngrok to solve it, but it would be super helpful if you can have a look.

Model name 'bart-large' was not found

When I tried to run the application I got the following error:

Traceback (most recent call last):
  File "app.py", line 4, in <module>
    import main
  File "/home/moussa/Documents/next_word_prediction/main.py", line 18, in <module>
    bart_tokenizer = BartTokenizer.from_pretrained('bart-large')
  File "/home/moussa/anaconda3/lib/python3.7/site-packages/transformers/tokenization_utils.py", line 911, in from_pretrained
    return cls._from_pretrained(*inputs, **kwargs)
  File "/home/moussa/anaconda3/lib/python3.7/site-packages/transformers/tokenization_utils.py", line 1014, in _from_pretrained
    list(cls.vocab_files_names.values()),
OSError: Model name 'bart-large' was not found in tokenizers model name list (facebook/bart-large, facebook/bart-large-mnli, facebook/bart-large-cnn, facebook/bart-large-xsum). We assumed 'bart-large' was a path, a model identifier, or url to a directory containing vocabulary files named ['vocab.json', 'merges.txt'] but couldn't find such vocabulary files at this path or url.