This is a replicate of DeepSeek-R1-Zero and DeepSeek-R1 training on small models with limited data

Fully open reproduction of DeepSeek-R1

Leveraging BERT and c-TF-IDF to create easily interpretable topics.

DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding

Unofficial implementation of Titans, SOTA memory for transformers, in Pytorch

Open Source alternative to Algolia and an Easier-to-Use alternative to ElasticSearch ⚡ 🔍 ✨ Fast, typo tolerant, in-memory fuzzy Search Engine for building delightful search experiences

Model for watermark classification implemented with PyTorch

Clustering for arbitrary data and dissimilarity function

Official Implementation of "HMT: Hierarchical Memory Transformer for Long Context Language Processing"

Embedding - Clustering Community Detection Framework

Sample code and notebooks for Generative AI on Google Cloud, including Gemini

A Curated List of Must-read Papers on Recommender System.

Scripts for fine-tuning Meta Llama3 with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supportin…

Finetune Llama 3, Mistral & Gemma LLMs 2-5x faster with 80% less memory

The official Meta Llama 3 GitHub site

Google Research

CoreNet: A library for training deep neural networks

Implementation of Nougat Neural Optical Understanding for Academic Documents

Cleora AI is a general-purpose model for efficient, scalable learning of stable and inductive entity embeddings for heterogeneous relational data.

MinHash, LSH, LSH Forest, Weighted MinHash, HyperLogLog, HyperLogLog++, LSH Ensemble

🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.

The Next Generation Multi-Modality Superintelligence

Facebook AI Research Sequence-to-Sequence Toolkit written in Python.

An open source implementation of CLIP.

Scenic: A Jax Library for Computer Vision Research and Beyond

Hrrformer: A Neuro-symbolic Self-attention Model (ICML23)

Automatically create Faiss knn indices with the most optimal similarity search parameters.