👑 Easy-to-use and powerful NLP and LLM library with 🤗 Awesome model zoo, supporting wide-range of NLP tasks from research to industrial applications, including 🗂Text Classification, 🔍 Neural Search, ❓ Question Answering, ℹ️ Information Extraction, 📄 Document Intelligence, 💌 Sentiment Analysis etc.

Accelerate local LLM inference and finetuning (LLaMA, Mistral, ChatGLM, Qwen, Baichuan, Mixtral, Gemma, Phi, etc.) on Intel CPU and GPU (e.g., local PC with iGPU, discrete GPU such as Arc, Flex and Max); seamlessly integrate with llama.cpp, Ollama, HuggingFace, LangChain, LlamaIndex, DeepSpeed, vLLM, FastChat, Axolotl, etc.

Unified embedding generation and search engine. Also available on cloud - cloud.marqo.ai

Profile PyTorch models for FLOPs and parameters, helping to evaluate computational efficiency and memory usage.

Implementation of Band Split Roformer, SOTA Attention network for music source separation out of ByteDance AI Labs

This is the official repository for the paper "Flora: Low-Rank Adapters Are Secretly Gradient Compressors".

My implementation of Machine Learning and Deep Learning papers from scratch.

A low-memory high-performance CPU-based API for Meta's No Language Left Behind (NLLB) using CTranslate2, hosted on Hugging Face Spaces.

Prompts, notebooks, and tools for generative pre-trained transformers.

prettier-plugin-jsonata

AI-First Process Automation with Large ([Language (LLMs) / Action (LAMs) / Multimodal (LMMs)] / Visual Language (VLMs)) Models

Transformer models for generative modeling of bioactive chemical small molecules

Signal Processing Fun (in the sun)

OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference

Testing the capabilities of the Llama 3 language model, specifically the Meta-Llama-3-8B-Instruct variant with 8 billion parameters.

Neural Search

A light-weight demo app for geospatial semantic search. Designed for text data with geospatial references.

Implementation of Alphafold 3 in Pytorch

Tools for understanding how transformer predictions are built layer-by-layer