FudanNLP/fastNLP
A lightweight NLP framework for Chinese text processing with built-in training loops, data processing utilities, and multi-framework backend support.
Not currently ranked — collecting fresh signals.
star history
fastNLP is a modular natural language processing framework developed at Fudan University designed to reduce engineering overhead in NLP projects. It provides convenient data processing with multi-process support, customizable training loops, and seamless switching between fp16 training, multi-GPU, and ZeRO optimization without code changes. The framework supports multiple deep learning frameworks as backends, primarily PyTorch, and includes loaders for Chinese NLP datasets like ChnSentiCorp.
Frequently asked
- What is FudanNLP/fastNLP?
- A lightweight NLP framework for Chinese text processing with built-in training loops, data processing utilities, and multi-framework backend support.
- Is fastNLP open source?
- Yes — FudanNLP/fastNLP is open source, released under the Apache-2.0 license.
- What language is fastNLP written in?
- FudanNLP/fastNLP is primarily written in Python.
- How popular is fastNLP?
- FudanNLP/fastNLP has 3.1k stars on GitHub.
- Where can I find fastNLP?
- FudanNLP/fastNLP is on GitHub at https://github.com/FudanNLP/fastNLP.