AVNLP

All

9 repositories

vectordb
Public
Created pipelines for Semantic Search, Metadata Filtering, Hybrid Search, Reranking, and Retrieval-Augmented Generation (RAG) for the TriviaQA, ARC, PopQA, FactScore, and Edgar datasets. These pipelines have been implemented for the Pinecone, Weaviate, Milvus and Chroma vector databases.
haystack pinecone weaviate milvus vector-database llm langchain
Python
•
MIT License
•0•3•0•0•Updated Jan 22, 2025Jan 22, 2025
llm-finetuning
Public
lora fine-tuning peft sft p-tuning qlora
Python
•
MIT License
•0•2•0•0•Updated Jan 19, 2025Jan 19, 2025
rag-pipelines
Public
pinecone rag crag llm self-rag
Python
•
MIT License
•0•2•0•0•Updated Jan 12, 2025Jan 12, 2025
llm-blender
Public
Ensembling LLMs using LLM-Blender
rag rankers llm-blender
Python
•
MIT License
•1•5•0•0•Updated Jan 9, 2025Jan 9, 2025
hyperparameter-tuning
Public
Effect of Optimizer Selection and Hyperparameter Tuning on Training Efficiency and LLM Performance
hyperparameter-tuning adam-optimizer sgd-optimizer sgd-momentum optimizers rmsprop-optimizer
Python
•
MIT License
•0•2•0•0•Updated Jan 9, 2025Jan 9, 2025
.github
Public
MIT License
•0•0•0•0•Updated Jan 2, 2025Jan 2, 2025
dataloaders
Public
Dataloaders is a versatile library designed for processing and formatting datasets to support various Retrieval-Augmented Generation (RAG) pipelines, facilitating efficient evaluation and analysis.
Python
•
MIT License
•0•2•0•0•Updated Dec 25, 2024Dec 25, 2024
rrf
Public
Performance Evaluation of Rankers and RRF Techniques for Retrieval Pipelines
ranker rrf reciprocal-rank-fusion similarity-ranker lost-in-the-middle-ranker diversity-ranker
Python
•
MIT License
•0•3•0•0•Updated Aug 21, 2024Aug 21, 2024
finbert-hf
Public
Python
•
MIT License
•0•0•0•0•Updated Jan 30, 2023Jan 30, 2023