Skip to content

Latest commit

 

History

History
100 lines (73 loc) · 2.65 KB

README.md

File metadata and controls

100 lines (73 loc) · 2.65 KB

Ps: we are adding transformer model soon

Text Gen 🐐

Downloads python tensorflow PyPI

Almost State-of-the-art Text Generation library

Text gen is a python library that allow you build a custom text generation model with ease 😄 Something sweet built with Tensorflow and Pytorch(coming soon) - This is the brain of Rosalove ai (https://rosalove.xyz/)

How to use it

Install text-gen

pip install -U text-gen

import the library

from text_gen import ten_textgen as ttg

Load your data. your data must be in a text format.

Download the example data from the example folder

load data

data = 'rl.csv'
text = ttg.loaddata(data)

build our Model Architeture

pipeline = ttg.tentext(text)
seq_text = pipeline.sequence(padding_method = 'pre')
configg = pipeline.configmodel(seq_text, lstmlayer = 128, activation = 'softmax', dropout = 0.25)

train model

model_history = pipeline.fit(loss = 'categorical_crossentropy', optimizer = 'adam', batch = 300, metrics = 'accuracy', epochs = 500, verbose = 0, patience = 10)

generate text using the phrase

pipeline.predict('hello love', word_length = 200, segment = True)

plot loss and accuracy

pipeline.plot_loss_accuracy()

Hyper parameter optimization

Tune your model to know the best optimizer, activation method to use.

pipeline.hyper_params(epochs = 500)
pipeline.saveModel('model')

use a saved model for prediction

#the corpus is the train text file
ttg.load_model_predict(corpus = corpus, padding_method = 'pre', modelname = '../input/model2/model2textgen.h5', sample_text = 'yo yo', word_length = 100)

Give us a star ⭐ 🐉

If you want to contribute, take a look at the issues and the Futurework.md file

Contributors