Search Results for "throughput" - Page 3
Sort By:
Showing 84 open source projects for "throughput"
View related business solutions-
Enterprise-grade ITSM, for every businessFreshservice is an intuitive, AI-powered platform that helps IT, operations, and business teams deliver exceptional service without the usual complexity. Automate repetitive tasks, resolve issues faster, and provide seamless support across the organization. From managing incidents and assets to driving smarter decisions, Freshservice makes it easy to stay efficient and scale with confidence.
-
Go From AI Idea to AI App FastAccess Gemini 3 and 200+ models. Build chatbots, agents, or custom models with built-in monitoring and scaling.
-
1
fairseq2
FAIR Sequence Modeling Toolkit 2
fairseq2 is a modern, modular sequence modeling framework developed by Meta AI Research as a complete redesign of the original fairseq library. Built from the ground up for scalability, composability, and research flexibility, fairseq2 supports a broad range of language, speech, and multimodal content generation tasks, including instruction fine-tuning, reinforcement learning from human feedback (RLHF), and large-scale multilingual modeling. Unlike the original fairseq—which evolved into a... -
2
DLRM
An implementation of a deep learning recommendation model (DLRM)
...It includes data loaders for standard benchmarks (like Criteo), training scripts, evaluation tools, and capabilities like mixed precision, gradient compression, and memory fusion to maximize throughput. -
3
xFormers
Hackable and optimized Transformers building blocks
xformers is a modular, performance-oriented library of transformer building blocks, designed to allow researchers and engineers to compose, experiment, and optimize transformer architectures more flexibly than monolithic frameworks. It abstracts components like attention layers, feedforward modules, normalization, and positional encoding, so you can mix and match or swap optimized kernels easily. One of its key goals is efficient attention: it supports dense, sparse, low-rank, and... -
4
Ring
Ring is a reasoning MoE LLM provided and open-sourced by InclusionAI
Ring is a reasoning Mixture-of-Experts (MoE) large language model (LLM) developed by inclusionAI. It is built from or derived from Ling. Its design emphasizes reasoning, efficiency, and modular expert activation. In its “flash” variant (Ring-flash-2.0), it optimizes inference by activating only a subset of experts. It applies reinforcement learning/reasoning optimization techniques. Its architectures and training approaches are tuned to enable efficient and capable reasoning performance.... -
$300 in Free Credit Towards Top Cloud ServicesStart your project in minutes. After credits run out, 20+ products include free monthly usage. Only pay when you're ready to scale.
-
5
DALI
A GPU-accelerated library containing highly optimized building blocks
...DALI addresses the problem of the CPU bottleneck by offloading data preprocessing to the GPU. Additionally, DALI relies on its own execution engine, built to maximize the throughput of the input pipeline. -
6
qvac-fabric-llm.cpp
QVAC Fabric: cross-platform LLM inference and fine-tuning
qvac-fabric-llm.cpp is a cross-platform large language model inference and fine-tuning engine built as an advanced fork of llama.cpp, designed to run efficiently across desktops, mobile devices, and heterogeneous GPU environments. The project focuses on removing hardware limitations traditionally associated with LLM deployment by enabling support for a wide range of backends, including Vulkan, Metal, CUDA, and CPU, making it accessible on devices ranging from smartphones to enterprise... -
7
Mooncake
Mooncake is the serving platform for Kimi
...Mooncake also introduces distributed key-value cache storage that allows inference systems to reuse previously computed attention states, significantly improving throughput in large-scale deployments. The system supports advanced networking technologies such as RDMA and NVMe over Fabric, enabling high-speed communication across clusters. -
8
Matrix
Multi-Agent daTa geneRation Infra and eXperimentation framework
Matrix is a distributed, large-scale engine for multi-agent synthetic data generation and experiments: it provides the infrastructure to run thousands of “agentic” workflows concurrently (e.g. multiple LLMs interacting, reasoning, generating content, data-processing pipelines) by leveraging distributed computing (like Ray + cluster management). The idea is to treat data generation as a “data-to-data” transformation: each input item defines a task, and the runtime orchestrates asynchronous,... -
9
marqo
Tensor search for humans
A tensor-based search and analytics engine that seamlessly integrates with your applications, websites, and workflows. Marqo is a versatile and robust search and analytics engine that can be integrated into any website or application. Due to horizontal scalability, Marqo provides lightning-fast query times, even with millions of documents. Marqo helps you configure deep-learning models like CLIP to pull semantic meaning from images. It can seamlessly handle image-to-image, image-to-text and... -
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
Synapse Machine Learning
Simple and distributed Machine Learning
...With the HTTP on Spark project, users can embed any web service into their SparkML models. For production-grade deployment, the Spark Serving project enables high throughput, sub-millisecond latency web services, backed by your Spark cluster. -
11
MiniMax-01
Large-language-model & vision-language-model based on Linear Attention
...MiniMax-Text-01 uses a hybrid attention architecture that blends Lightning Attention, standard softmax attention, and Mixture-of-Experts (MoE) routing to achieve both high throughput and long-context reasoning. It has 456 billion total parameters with 45.9 billion activated per token and is trained with advanced parallel strategies such as LASP+, varlen ring attention, and Expert Tensor Parallelism, enabling a training context of 1 million tokens and up to 4 million tokens at inference. MiniMax-VL-01 extends this core by adding a 303M-parameter Vision Transformer and a two-layer MLP projector in a ViT–MLP–LLM framework, allowing the model to process images at dynamic resolutions up to 2016×2016. -
12
DeiT (Data-efficient Image Transformers)
Official DeiT repository
...Its key idea is a specialized distillation strategy—including a learnable “distillation token”—that lets a transformer learn effectively from a CNN or transformer teacher on modest-scale datasets. The project provides compact ViT variants (Tiny/Small/Base) that achieve excellent accuracy–throughput trade-offs, making transformers practical beyond massive pretraining regimes. Training involves carefully tuned augmentations, regularization, and optimization schedules to stabilize learning and improve sample efficiency. The repo offers pretrained checkpoints, reference scripts, and ablation studies that clarify which ingredients matter most for data-efficient ViT training. -
13
Hiera
A fast, powerful, and simple hierarchical vision transformer
Hiera is a hierarchical vision transformer designed to be fast, simple, and strong across image and video recognition tasks. The core idea is to use straightforward hierarchical attention with a minimal set of architectural “bells and whistles,” achieving competitive or superior accuracy while being markedly faster at inference and often faster to train. The repository provides installation options (from source or Torch Hub), a model zoo with pre-trained checkpoints, and code for evaluation... -
14
TensorHouse
A collection of reference Jupyter notebooks and demo AI/ML application
TensorHouse is a scalable reinforcement learning (RL) platform that focuses on high-throughput experience generation and distributed training. It is designed to efficiently train agents across multiple environments and compute resources. TensorHouse enables flexible experiment management, making it suitable for large-scale RL experiments in both research and applied settings. -
15
OnnxStream
Lightweight inference library for ONNX files, written in C++
...The recommended minimum RAM/VRAM for Stable Diffusion 1.5 is typically 8GB. Generally, major machine learning frameworks and libraries are focused on minimizing inference latency and/or maximizing throughput, all of which at the cost of RAM usage. So I decided to write a super small and hackable inference library specifically focused on minimizing memory consumption: OnnxStream. OnnxStream is based on the idea of decoupling the inference engine from the component responsible for providing the model weights, which is a class derived from WeightsProvider. ... -
16
Punica
Serving multiple LoRA finetuned LLM as one
...The system includes specialized CUDA kernels that enable batched GPU operations across different LoRA models simultaneously. This design allows a single GPU cluster to host many task-specific models while maintaining high throughput and minimal latency. The architecture also includes scheduling mechanisms that coordinate requests from multiple tenants and distribute workloads efficiently across available resources. -
17
LLaMA-MoE
Building Mixture-of-Experts from LLaMA with Continual Pre-training
LLaMA-MoE is an open-source project that builds mixture-of-experts language models from LLaMA through expert partitioning and continual pre-training. The repository is centered on making MoE research more accessible by offering smaller and more affordable models with only about 3.0 to 3.5 billion activated parameters, which helps reduce deployment and experimentation costs. Its architecture works by splitting LLaMA feed-forward networks into sparse experts and adding gating mechanisms so... -
18
Medusa
Framework for Accelerating LLM Generation with Multiple Decoding Heads
Medusa is a framework aimed at accelerating the generation capabilities of Large Language Models (LLMs) by employing multiple decoding heads. This approach allows for parallel processing during text generation, significantly enhancing throughput and reducing response times. Medusa is designed to be simple to implement and integrates with existing LLM infrastructures, making it a practical solution for scaling LLM applications. -
19
FastViT
This repository contains the official implementation of research
FastViT is an efficient vision backbone family that blends convolutional inductive biases with transformer capacity to deliver strong accuracy at mobile and real-time inference budgets. Its design pursues a favorable latency-accuracy Pareto curve, targeting edge devices and server scenarios where throughput and tail latency matter. The models use lightweight attention and carefully engineered blocks to minimize token mixing costs while preserving representation power. Training and inference recipes highlight straightforward integration into common vision tasks such as classification, detection, and segmentation. The codebase provides reference implementations and checkpoints that make it easy to evaluate or fine-tune on downstream datasets. ... -
20
FFCV
Fast Forward Computer Vision (and other ML workloads!)
ffcv is a drop-in data loading system that dramatically increases data throughput in model training. From gridding to benchmarking to fast research iteration, there are many reasons to want faster model training. Below we present premade codebases for training on ImageNet and CIFAR, including both (a) extensible codebases and (b) numerous premade training configurations. -
21
FasterTransformer
Transformer related optimization, including BERT, GPT
FasterTransformer is a high-performance inference library designed to accelerate transformer-based models such as BERT, GPT, and T5 on NVIDIA GPUs. It provides optimized implementations of transformer encoder and decoder layers using CUDA, cuBLAS, and custom kernels to maximize throughput and minimize latency. The library supports multiple deep learning frameworks, including TensorFlow, PyTorch, and Triton, allowing developers to integrate it into existing pipelines without major changes. It includes advanced optimization techniques such as mixed precision, tensor parallelism, and efficient memory management, enabling large models to run across multiple GPUs and nodes. ... -
22
LightSeq
A High Performance Library for Sequence Processing and Generation
Lightseq is a high-performance library focused on efficient inference and training for deep learning models, especially large language models (LLMs) and transformer-based architectures. Its goal is to optimize both memory usage and computational throughput, enabling faster training or inference on limited hardware while maintaining model quality. Lightseq provides optimized CUDA kernels, quantization strategies, and runtime optimizations tailored for transformer operations — which often are bottlenecks in conventional frameworks — thereby reducing memory footprint, improving speed, and making deployment of large-scale models more accessible. ... -
23
OmicSelector
Feature selection and deep learning modeling for omic biomarker study
OmicSelector is an environment, Docker-based web application, and R package for biomarker signature selection (feature selection) from high-throughput experiments and others. It was initially developed for miRNA-seq (small RNA, smRNA-seq; hence the name was miRNAselector), RNA-seq and qPCR, but can be applied for every problem where numeric features should be selected to counteract overfitting of the models. Using our tool, you can choose features, like miRNAs, with the most significant diagnostic potential (based on the results of miRNA-seq, for validation in qPCR experiments). -
24
hora
Efficient approximate nearest neighbor search algorithm collections
...The library is written in Rust and emphasizes performance, safety, and efficient memory management, making it suitable for production-grade applications requiring low latency and high throughput. -
25
TurboTransformers
Fast and user-friendly runtime for transformer inference
TurboTransformers is a high-performance inference framework optimized for running Transformer models efficiently on CPUs and GPUs. It improves latency and throughput for NLP applications.
