Search Results for "speech processing"

spaCy is a library built on the very latest research for advanced Natural Language Processing (NLP) in Python and Cython. Since its inception it was designed to be used for real world applications-- for building real products and gathering real insights. It comes with pretrained statistical models and word vectors, convolutional neural network models, easy deep learning integration and so much more. spaCy is the fastest syntactic parser in the world according to independent benchmarks, with...

Stanza is a collection of accurate and efficient tools for the linguistic analysis of many human languages. Starting from raw text to syntactic analysis and entity recognition, Stanza brings state-of-the-art NLP models to languages of your choosing. Stanza is a Python natural language analysis package. It contains tools, which can be used in a pipeline, to convert a string containing human language text into lists of sentences and words, to generate base forms of those words, their parts of...

SVoice is a PyTorch-based implementation of Facebook Research’s study on speaker voice separation as described in the paper “Voice Separation with an Unknown Number of Multiple Speakers.” This project presents a deep learning framework capable of separating mixed audio sequences where several people speak simultaneously, without prior knowledge of how many speakers are present. The model employs gated neural networks with recurrent processing blocks that disentangle voices over multiple...