Search Results for "distributed shared memory"
Sort By:
Showing 74 open source projects for "distributed shared memory"
View related business solutions-
$300 Free Credits to Build on Google CloudPut your $300 in credit toward real workloads, then keep building with free monthly usage for 20+ products. No commitment and no charge until you upgrade.
-
Custom VMs From 1 to 96 vCPUs With 99.95% UptimeLive migration and automatic failover keep workloads online through maintenance. One free e2-micro VM every month.
-
1
MSA: Memory Sparse Attention
Trainable latent-memory framework for 100M-token contexts
MSA, or Memory Sparse Attention, is a research framework for scaling language-model memory to extremely long contexts. It replaces full attention over all tokens with sparse selection of compressed latent memory states. Document-wise rotary position encoding and top-k routing keep training and inference close to linear complexity. A tiered KV-cache design stores routing keys on GPU while larger content states can remain on CPU. Its Memory Parallel engine distributes scoring and transfers... -
2
I hate money
A simple shared budget manager web application
I hate money is a web application made to ease shared budget management. It keeps track of who bought what, when, and for whom; and helps to settle the bills. I hate money is written in python, using the flask framework. It’s developed with ease of use in mind and is trying to keep things simple. Hope you (will) like it! The code is distributed under a BSD beerware derivative: if you meet the people in person and you want to pay them a craft beer, you are highly encouraged to do so. -
3
OpenSquilla
Token-Efficient AI Agent with same budget, higher intelligence density
OpenSquilla is a token-efficient microkernel AI agent runtime designed for CLI, web UI, and chat-based workflows. It routes each turn through a shared loop that can select lower-cost models when appropriate while preserving tool dispatch, retries, memory, and decision logging. The project supports multiple LLM providers through a pluggable provider layer, making it adaptable to different model ecosystems. It includes persistent memory, built-in web search, on-device embeddings, and sandboxing for safer execution. ... -
4
MemClaw
Persistent memory for AI agent fleets (OSS)
MemClaw is an open-source governed shared memory platform for AI agent fleets. It is designed to help agents remember information across sessions, teams, tools, and models instead of keeping knowledge trapped inside isolated conversations. The project emphasizes enterprise-style governance, including permissions, tenant isolation, audit trails, visibility scopes, and agent trust tiers. -
Host LLMs in Production With On-Demand GPUsDeploy your model, get an endpoint, pay only for compute time. No GPU provisioning or infrastructure management required.
-
5
EverOS
One portable memory layer for every AI agent
...The system is designed to preserve user context and agent experience across sessions without forcing dependence on a hosted service. EverOS can serve as shared memory infrastructure for coding assistants, personal agents, and custom applications. Its architecture emphasizes user ownership, portability, and self-evolving reuse of accumulated knowledge. -
6
Former.skill
AI skill generator designed to reconstruct conversational personality
Ex Skill is an AI skill generator designed to reconstruct the conversational personality and shared memories of a former partner from personal data. It can analyze exported chats, photos, social media records, PDFs, images, and manually entered descriptions. The system builds separate memory and persona layers covering relationship history, habits, emotional patterns, communication style, and behavior. Generated personas can respond in a style modeled after the source material. ... -
7
vLLM
A high-throughput and memory-efficient inference and serving engine
vLLM is a fast and easy-to-use library for LLM inference and serving. High-throughput serving with various decoding algorithms, including parallel sampling, beam search, and more. -
8
Colossal-AI
Making large AI models cheaper, faster and more accessible
The Transformer architecture has improved the performance of deep learning models in domains such as Computer Vision and Natural Language Processing. Together with better performance come larger model sizes. This imposes challenges to the memory wall of the current accelerator hardware such as GPU. It is never ideal to train large models such as Vision Transformer, BERT, and GPT on a single GPU or a single machine. There is an urgent demand to train models in a distributed environment. However, distributed training, especially model parallelism, often requires domain expertise in computer systems and architecture. ... -
9
Qwen
The official repo of Qwen chat & pretrained large language model
Qwen is a series of large language models developed by Alibaba Cloud, consisting of various pretrained versions like Qwen-1.8B, Qwen-7B, Qwen-14B, and Qwen-72B. These models, which range from smaller to larger configurations, are designed for a wide range of natural language processing tasks. They are openly available for research and commercial use, with Qwen's code and model weights shared on GitHub. Qwen's capabilities include text generation, comprehension, and conversation, making it a... -
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
10
Olmo-core
PyTorch building blocks for the OLMo ecosystem
OLMo-core is a Python library from AI2 that provides reusable PyTorch building blocks for training large language models at scale. It underpins the broader OLMo ecosystem and includes components for models, data loading, distributed execution, optimization, checkpointing, evaluation, and generation. Official scripts reproduce training configurations for released OLMo models. The library supports advanced attention backends, mixture-of-experts architectures, float8 training, and fused low-memory operations through optional dependencies. Jobs can run with standard torchrun or AI2’s Beaker infrastructure. ... -
11
Claude-Flow
The leading agent orchestration platform for Claude
...It enables developers to coordinate multiple specialized AI agents in real time through a hive-mind architecture, combining swarm intelligence, neural reasoning, and a powerful set of 87 Modular Control Protocol (MCP) tools. The platform supports both quick swarm tasks and persistent multi-agent sessions known as hives, facilitating distributed AI collaboration with persistent contextual memory. At its core, Claude-Flow integrates Dynamic Agent Architecture (DAA) for self-organizing agent management, neural pattern recognition accelerated by WebAssembly SIMD, and a SQLite-based memory system for context retention and knowledge persistence across tasks. It automates development workflows via pre- and post-operation hooks, providing seamless coordination, code formatting, validation, and performance optimization. -
12
DeepSeed
Deep learning optimization library making distributed training easy
DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective. DeepSpeed delivers extreme-scale model training for everyone, from data scientists training on massive supercomputers to those training on low-end clusters or even on a single GPU. Using current generation of GPU clusters with hundreds of devices, 3D parallelism of DeepSpeed can efficiently train deep learning models with trillions of parameters. With just a single GPU,... -
13
METATRON
AI-powered penetration testing assistant using local LLM on linux
METATRON is a multi-agent AI orchestration framework designed to coordinate complex workflows across multiple intelligent agents. It provides a structured system for task delegation, communication, and collaboration between agents. The framework emphasizes scalability, allowing multiple agents to work together on large or complex problems. It includes mechanisms for managing context, memory, and execution flow across tasks. METATRON is particularly useful for building advanced AI systems... -
14
Dask
Parallel computing with task scheduling
Dask is a Python library for parallel and distributed computing, designed to scale analytics workloads from single machines to large clusters. It integrates with familiar tools like NumPy, Pandas, and scikit-learn while enabling execution across cores or nodes with minimal code changes. Dask excels at handling large datasets that don’t fit into memory and is widely used in data science, machine learning, and big data pipelines. -
15
OSS-Fuzz
OSS-Fuzz - continuous fuzzing for open source software
OSS-Fuzz is a large-scale fuzz testing platform developed by Google to improve the security and reliability of widely used open source software. Fuzz testing is a proven method for uncovering programming errors such as buffer overflows and memory leaks, which can lead to severe security vulnerabilities. By leveraging guided in-process fuzzing, Google has already identified thousands of issues in projects like Chrome, and this initiative extends the same capabilities to the broader open source community. OSS-Fuzz integrates modern fuzzing engines with sanitizers and runs them at scale in a distributed environment, providing automated testing and continuous monitoring. ... -
16
FramePack
Lets make video diffusion practical
FramePack explores compact representations for sequences of image frames, targeting tasks where many near-duplicate frames carry redundant information. The idea is to “pack” frames by detecting shared structure and storing differences efficiently, which can accelerate training or inference on video-like data. By reducing I/O and memory bandwidth, datasets become lighter to load while models still see the essential temporal variation. The repository demonstrates both packing and unpacking steps, making it straightforward to integrate into preprocessing pipelines. ... -
17
Lingvo
Framework for building neural networks
Lingvo is a TensorFlow based framework focused on building and training sequence models, especially for language and speech tasks. It was originally developed for internal research and later open sourced to support reproducible experiments and shared model implementations. The framework provides a structured way to define models, input pipelines, and training configurations using a common interface for layers, which encourages reuse across different tasks. It has been used to implement state... -
18
MLX LM
Run LLMs with MLX
...Users can perform low-rank and full-model fine-tuning, including training quantized models. Its Python API and command-line tools support streaming generation, configurable sampling, and interactive chat. Additional capabilities include distributed inference, prompt caching, and memory-efficient processing of long contexts. -
19
huey
A little task queue for python
...It gives developers a clean API for running background jobs outside the main request or execution flow. The project supports several storage backends, including Redis, Valkey, Redict, SQLite, the file system, and in-memory storage. It can execute tasks with processes, threads, or greenlets, which makes it adaptable to different workloads. Huey also supports scheduled tasks, recurring tasks, retries, task priorities, result storage, expiration, locking, rate limits, timeouts, pipelines, groups, and chords. Overall, it is a practical option for Python teams that want background processing without the complexity of a heavier distributed task system. -
20
Smallpond
A lightweight data processing framework built on DuckDB and 3FS
smallpond is a lightweight distributed data processing framework built by DeepSeek, designed to scale DuckDB workloads over clusters using their 3FS (Fire-Flyer File System) backend. The idea is to preserve DuckDB’s fast analytics engine but lift it from single-node to multi-node settings, giving you the ability to operate on large datasets (e.g. petabyte scale) without moving to a heavyweight system like Spark. Users write Python-like code (via DataFrame APIs or SQL strings) to express... -
21
Xtuner
A Next-Generation Training Engine Built for Ultra-Large MoE Models
...Its architecture incorporates memory-efficient optimizations that allow researchers to train large models even when computational resources are limited. XTuner is also designed to integrate with modern AI ecosystems, supporting multimodal training, reinforcement learning optimization, and instruction tuning pipelines. -
22
CogVideo
Text and image to video generation: CogVideoX and CogVideo
CogVideo is an open-source family of advanced video generation models that can create videos from text, images, or existing video inputs. Built on large-scale Transformer and diffusion architectures, it enables multimodal generation across text-to-video, image-to-video, and video continuation tasks. The latest CogVideoX models offer higher resolution outputs, longer video durations, and improved controllability through prompt engineering. The project includes tools for inference,... -
23
CUDA Python
Performance meets Productivity
...It integrates tightly with the broader Python GPU ecosystem, including Numba for kernel compilation and CCCL for parallel primitives, allowing developers to write performant code without leaving Python. The toolkit also includes utilities for profiling, memory management, distributed computing, and numerical operations, making it suitable for scientific computing, AI, and data processing workloads. -
24
DLRM
An implementation of a deep learning recommendation model (DLRM)
DLRM (Deep Learning Recommendation Model) is Meta’s open-source reference implementation for large-scale recommendation systems built to handle extremely high-dimensional sparse features and embedding tables. The architecture combines dense (MLP) and sparse (embedding) branches, then interacts features via dot product or feature interactions before passing through further dense layers to predict click-through, ranking scores, or conversion probabilities. The implementation is optimized for... -
25
J-Space Cognition Suite V3.6
AI cognitive-enhancement Skills based on Anthropic's J-space
J-Space Cognition Suite is a model-agnostic inference-time control system for improving deep reasoning, long-horizon work, tool use, verification, and recovery in AI agents. It is distributed as a cross-platform Skill that leaves model weights and training unchanged. The suite manages an agent's accessible working representations through selective loading instead of applying every mechanism to every task. Its fast, full, and loop modes scale from simple checks to multi-stage work requiring persistent state. Core mechanisms include shared workspace anchors, compact reasoning tracks, metacognitive control, explicit intermediate reasoning, and empirical verification. ...
