minimaxir/textgenrnn
Easily train your own text-generating neural network of any size and complexity on any text dataset with a few lines of code.
Built on Keras/TensorFlow, textgenrnn implements character- or word-level RNNs with modern architectural improvements including attention-weighting, skip connections, and optional contextual label training to accelerate convergence. The framework leverages CuDNN-optimized LSTM implementations for GPU training while supporting CPU inference, and includes pretrained weights (2 MB) derived from diverse Reddit data that can be fine-tuned on new datasets with momentum-based optimization and learning rate decay to prevent gradient instability.
4,931 stars and 204 monthly downloads. No commits in the last 6 months. Available on PyPI.
Stars
4,931
Forks
738
Language
Python
License
—
Category
Last pushed
Jul 17, 2022
Monthly downloads
204
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/minimaxir/textgenrnn"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
vpj/python_autocomplete
A simple neural network for python autocompletion
jackcook/predictive-spy
Spying on Apple’s new predictive text model
yuazi/keyboard-AI
Self-learning keyboard layout optimizer using n-gram statistics and an ergonomic cost model
realAltamashAli/bilingual-swiftkey-nlp
A high-performance bilingual (EN/DE) text prediction engine using Katz Back-off N-gram modeling,...
subhadipsinha722133/Spelling-Checker-Web-App
📝 A simple and intuitive web application built with Streamlit and pyspellchecker to help you...