jmete / riverbed

Tools for content datamining and NLP at scale

riverbed

Tools for content datamining and NLP at scale.

motiviation

Given a set of content including video, image and/or text content in human language, or code, we would like to:

• create interesting features for the content, including augmenting the content.
• cluster, explore, search and visualize the content.
• label and create classifiers for the content.
• search, filter, store and share the content to user and to other AI models

The content may be domain specific and may change over time. If a change is significant, we would like to be notified of the changes and/or automatically re-run some or all of the above.

installation

git clone https://github.com/ontocord/riverbed/
chmod ugo+x /content/riverbed/bin/lmplz
pip install https://github.com/kpu/kenlm/archive/master.zip
pip install dataset datasets fasttext indexed_gzip whoosh transformers sentencepiece spacy nltk fast-pytorch-kmeans mmh3 tqdm
pip install tsnecuda==3.0.1+cu112 -f https://tsnecuda.isx.ai/tsnecuda_stable.html
python -m spacy download en_core_web_md
python -m nltk.downloader stopwords

history

Originally written by Ontocord, LLC. Donated to LAION for the open source community.