Distributed task queue with full async support

An app that brings language models directly to your phone.

Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.

The official repo of Qwen (通义千问) chat & pretrained large language model proposed by Alibaba Cloud.

Fast and memory-efficient exact attention

Examples in the MLX framework

[ICML2024 (Oral)] Official PyTorch implementation of DoRA: Weight-Decomposed Low-Rank Adaptation

YaFSDP: Yet another Fully Sharded Data Parallel

Simple Python library/structure to ablate features in LLMs which are supported by TransformerLens

Solves basic Russian NLP tasks, API for lower level Natasha projects

MLX: An array framework for Apple silicon

🦁 Lion, new optimizer discovered by Google Brain using genetic algorithms that is purportedly better than Adam(w), in Pytorch

Python scripts performing object detection using the YOLOv8 model in ONNX.

Kolmogorov Arnold Networks