A fast, efficient universal vector embedding utility package.

Large Context Attention

Implementation of a memory efficient multi-head attention as proposed in the paper, "Self-attention Does Not Need O(n²) Memory"

Easy Parallel Library (EPL) is a general and efficient deep learning framework for distributed model training.

[CVPR 2025 Highlight] The official CLIP training codebase of Inf-CL: "Breaking the Memory Barrier: Near Infinite Batch Size Scaling for Contrastive Loss". A super memory-efficiency CLIP training scheme.

APOLLO: SGD-like Memory, AdamW-level Performance; MLSys'25 Oustanding Paper Honorable Mention

Train Dense Passage Retriever (DPR) with a single GPU

Code for: "Social-Implicit: Rethinking Trajectory Prediction Evaluation and The Effectiveness of Implicit Maximum Likelihood Estimation" Accepted @ ECCV2022

Toolkit for highly memory efficient analysis of single-cell RNA-Seq, scATAC-Seq and CITE-Seq data. Analyze atlas scale datasets with millions of cells on laptop.

[ACL 2023] The official implementation of "CAME: Confidence-guided Adaptive Memory Optimization"

IISAN: Efficiently Adapting Multimodal Representation for Sequential Recommendation with Decoupled PEFT

Lowering PyTorch's Memory Consumption for Selective Differentiation

The official implementation of Memory-efficient DQN algorithm.

Scalable Parameter and Memory Efficient Pretraining for LLM: Recent Algorithmic Advances and Benchmarking

A fast and memory efficient way to load large CSV files (Timeseries data) in Pandas.

Memory-efficient probabilistic counter namely Morris Counter

A lightweight and memory-efficient HTTP client module for MicroPython, optimized for use cases on resource-constrained devices such as ESP32.

Structured State Matrix Architecture (SSMA) is a high-performance framework designed for efficient sequence modeling, combining structured state space models with adaptive attention mechanisms.

black-forest-labs/FLUX.1-dev and black-forest-labs/FLUX.1-Krea-dev. This merged model integrates the capabilities of the rectified flow transformer FLUX.1-dev, known for competitive prompt following and high-quality outputs, with FLUX.1-Krea-dev, a guidance distilled model emphasizing aesthetics and photorealism.

Quantized 2D Gaussian Splatting for memory efficient 3D Reconstruction and Novel View Synthesis