alejandrods / cover-letter-generator-gpt2

Service for generating cover letter using GPT-2 model, deployed on Google Cloud.

Hi, my name is Alex 👋

My website:

I'm a Machine Learning Engineer, Aeronautical Engineer, Teacher and Climber!

• 📚 Machine Learning Engineer holding a Master's degree by the University of Sydney.
• 🔭 I’m taking a course of Redis
• ✍ I’m looking to write more articles on Medium
• ⚡ Fun fact: I love to climb and travel and I'm currently living in Australia ☀️

Connect with me:

Languages and Tools:

---

📕 Latest Projects

• Analysis of robust classification algorithms for overcoming class-dependant labelling noise: Forward, Importance Re-weighting and T-revision - Code! 💻
• Analysis of the robustness of non-negative matrix factorization (NMF) techniques: L2-norm, L1-norm, and L2,1-norm - Code! 💻
• Generating cover letters with the GPT-2 model - Deployed on Google Cloud - Try it here! 🕹️
• Named Entity Recognition in travel-related documents using RNN - Deployed on Google Cloud - Try it here! ✈️
• Call Center to analyze automatically if a person can be infected with Coronavirus - AWS Connect + AWS Lex Bot - Call me! - [phone] 📞
• Service for generating text with the GPT-2 model - Deployed on Google Cloud - Try it here! 🖨️
• Alexa Skill using Lambda Functions and Web Scraping - Try it here! 🤖

---

📕 Latest Blog Posts

• A little bit of strange/interesting Datasets for Machine Learning
• Measuring social distance in the time of Covid-19
• Noise: It’s not always annoying
• Style Transfer Pytorch
• Neural Network from Scratch

cover-letter-generator-gpt2's Issues

Google Colab

Do you know how I can run this on Google Colab?

`# -*- coding: utf-8 -*-

import os
import json
import flask
import logging
from textblob import TextBlob
from generate import generate
from flask_cors import CORS, cross_origin
from flask_ngrok import run_with_ngrok
from flask import Flask, render_template, url_for
from flask_wtf import FlaskForm
from wtforms import StringField, SubmitField, TextAreaField
from wtforms.validators import DataRequired
import torch
from transformers import GPT2LMHeadModel, GPT2Tokenizer
from flask import Flask, request, render_template
import json


app = flask.Flask(__name__)
CORS(app)
app.config['PROPAGATE_EXCEPTIONS'] = True
app = Flask(__name__, root_path='/content/cover-letter-generator-gpt2')
app.config['SECRET_KEY'] = '9bad6913d4358ac1395c5c94370ed090'
run_with_ngrok(app)
print(app.root_path)

# Logs
logging.basicConfig(level=logging.INFO,
                    format='%(asctime)s :: %(levelname)s :: %(message)s')


# Load Environment Variables
ROUTE_GENERATE = os.environ.get('ROUTE_GENERATE', 'generate')
N_TOKENS = int(os.environ.get('N_TOKENS', 20))

# This create a json file with the environment variables for javascript script
data_env = {
    "APP_URL": "{}:{}".format(HOST, PORT) if 'localhost' in HOST
    else "{}".format(HOST),
    "ROUTE": "/{}".format(ROUTE_GENERATE),
}

# Save env.json file
with open(os.path.join("/content/cover-letter-generator-gpt2/src/static", "env.json"), 'w') as f:
    json.dump(data_env, f)


@app.route('/content/cover-letter-generator-gpt2/src/templates/main.html')
@cross_origin()
def index():
    """
    Function to render the html file

    :return:
    """

    return flask.render_template('main.html')


@app.route('/generate', methods=['POST'])
@cross_origin()
def prediction():
    """
    Function to receive an user query and return generated text

    :return: Generated text
    """

    income_query = json.loads(flask.request.data)['input']

    logging.info('Query received: {}'.format(income_query))

    arg_lang = TextBlob(income_query)
    lang = arg_lang.detect_language()
    logging.info('Lang detected: {}'.format(lang))

    if lang == 'en':
        res = generate(income_query, size=N_TOKENS)
        generated = res.split(income_query)[-1]
    else:
        res = "lang_det_err"
        generated = "lang_det_err"

    logging.info('Result: {}'.format(res))

    return flask.jsonify({
        'version': 'v1.0.0',
        'body': {'query': income_query,
                 'generated': generated,
                 'result': res}
        }), 200


@app.route('/health_liveness')
@cross_origin()
def health_check():
    """
    Check status code

    :return:
    """

    return flask.jsonify({'status': 'success'}), 200


if __name__ == "__main__":
    app.run()
`