dorenwick Goto Github PK

atomic_roberta

A repo for training a Roberta language model base architecture, atomically sized because the models have sub 10million parameters.

cloudvectordb

A script for running on very powerful cloud computer, for building a very large dataset of triplets, then training encoders, then building the embeddings with the encoder, then building the vectordb with the encoder.

datasetsmergesplit

Some wrappers on Huggingface for complex dataset splitting and merging operations.

faiss-metadata-dense

This fork of faiss will seek to incorporate multi-vector and dense-vector-retrieval into the library.

gantrithor_installer

This repo is for constructing installers for the gantrithor

honorsthesis

sentence_encoder_distillation

This is a script for distilling the following model = SentenceTransformer("all-MiniLM-L6-v2") into smaller dimension models.

spanmarkerner_hf

SpanMarker for Named Entity Recognition, HuggingFace version