Search Results for "most" - Page 7
Sort By:
Showing 627 open source projects for "most"
View related business solutions-
Build Agents and Models on One PlatformGemini Enterprise Agent Platform is Google Cloud's comprehensive platform for developers to build, scale, govern, and optimize agents and models. Choose from Google's most advanced models and third-party models like Anthropic's Claude Model Family.
-
Demo Series - Small Business Backup By VeeamWatch this on-demand demo series and learn how to protect your Microsoft 365 data with clear, simple, actionable steps that are easy to implement for businesses of all sizes.
-
1
Lightweight' GAN
Implementation of 'lightweight' GAN, proposed in ICLR 2021
...You can test and see how your images will be augmented before they pass into a neural network (if you use augmentation). The general recommendation is to use suitable augs for your data and as many as possible, then after some time of training disable the most destructive (for image) augs. You can turn on automatic mixed precision with one flag --amp. You should expect it to be 33% faster and save up to 40% memory. Aim is an open-source experiment tracker that logs your training runs, and enables a beautiful UI to compare them. -
2
VOID
Video Object and Interaction Deletion
...Built on top of transformer-based architectures and fine-tuned for video inpainting tasks, the system uses interaction-aware mask conditioning to ensure temporal consistency across frames. One of its most notable capabilities is its ability to simulate realistic scene behavior after object removal, such as causing an object to fall naturally if its support is removed, which significantly enhances realism. -
3
SageAttention
NeurIPS2025 Spotlight] Quantized Attention
SageAttention is an open-source optimization library designed to accelerate the attention mechanism used in transformer-based neural networks. Since attention operations are often the most computationally expensive component of modern AI models, SageAttention introduces quantization techniques that significantly reduce computational overhead while preserving model accuracy. The system achieves this by using low-precision numerical formats such as INT4, FP8, or INT8 to represent key matrices within the attention computation. ... -
4
Quod Libet
Music player and music library manager for Linux, Windows, and macOS
...It lets you display and edit any tags you want in the file, for all the file formats it supports. Unlike some, Quod Libet will scale to libraries with tens of thousands of songs. It also supports most of the features you’d expect from a modern media player. -
99.99% Uptime for MySQL and PostgreSQL DatabasesCloud SQL Enterprise Plus delivers near-zero downtime with 35 days of point-in-time recovery. Supports MySQL, PostgreSQL, and SQL Server.
-
5
Jittor
Jittor is a high-performance deep learning framework
...Jittor also contains a wealth of high-performance model libraries, including image recognition, detection, segmentation, generation, differentiable rendering, geometric learning, reinforcement learning, etc. The front-end language is Python. Module Design and Dynamic Graph Execution is used in the front-end, which is the most popular design for deep learning framework interface. The back-end is implemented by high-performance languages, such as CUDA, C++. Jittor'op is similar to NumPy. Let's try some operations. We create Var a and b via operation jt.float32, and add them. Printing those variables shows they have the same shape and dtype. -
6
Mezzanine
CMS framework for Django
...In many ways it is like many other content management tools, offering an intuitive interface for managing all of your content. But Mezzanine is different in that it provides most of its functionality by default. While other platforms rely heavily on modules or reusable applications, Mezzanine comes ready with all the functionality you need, making it the more efficient choice. Mezzanine has a simple yet highly extensible architecture that lets you really get into the code. Apart from the features that come with Django such as MVC architecture, ORM, templating and caching, Mezzanine comes with a great many other features. ... -
7
PS2 Cover
PS2 Covers Collection
...The project also includes tools and instructions for integrating the covers into emulator workflows, enabling seamless automation. Its scale and completeness make it one of the most comprehensive resources for retro gaming visuals. Overall, ps2-covers enhances the user experience of emulation by adding organized and accessible visual metadata. -
8
MiniCPM-o
A GPT-4o Level MLLM for Vision, Speech and Multimodal Live Streaming
...Capable of running on end-side devices such as smartphones and tablets, it provides powerful features like real-time speech conversation, video understanding, and multimodal live streaming. With 8 billion parameters, MiniCPM-o 2.6 surpasses its predecessors in versatility and efficiency, making it one of the most robust models available. It supports both text and audio inputs to generate outputs in various forms, including voice cloning, emotion control, and interactive role-playing. -
9
Opacus
Training PyTorch models with differential privacy
...It supports training with minimal code changes required on the client, has little impact on training performance, and allows the client to online track the privacy budget expended at any given moment. Vectorized per-sample gradient computation that is 10x faster than micro batching. Supports most types of PyTorch models and can be used with minimal modification to the original neural network. Open source, modular API for differential privacy research. Everyone is welcome to contribute. ML practitioners will find this to be a gentle introduction to training a model with differential privacy as it requires minimal code changes. ... -
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
git-delete-merged-branches
Command-line tool to delete merged Git branches
...Provider agnostic: Works with GitHub, GitLab, Gitea and any other Git hosting. Takes safety seriously. Deletion is a sharp knife that requires care. While git reflog would have your back in most cases, git-delete-merged-branches takes safety seriously. git push is used with --force-with-lease so if the server and you have a different understanding of that branch, it is not deleted. There is no use of os.system or shell code to go wrong. -
11
Sentry
Cross-platform application monitoring and error tracking software
Sentry is a cross-platform, self-hosted error monitoring solution that helps software teams discover, monitor and fix errors in real-time. The most users and logs will have to provide are the clues, and Sentry provides the answers. Sentry offers enhanced application performance monitoring through information-laden stack traces. It lets you build better software faster and more efficiently by showing you all issues in one place and providing the trail of events that lead to errors. ... -
12
Train LLM From Scratch
A straightforward method for training your LLM
...It supports training smaller or larger models, including million- and billion-parameter configurations depending on available hardware. A major goal is accessibility, since the author frames it as possible to train models using a single GPU. It is most useful for learners, researchers, and developers who want practical exposure to LLM internals. -
13
Viral-Clips-Crew
Your CrewAI Powered Video Editing Assistant
Viral-Clips-Crew is an AI-driven video processing pipeline designed to generate short-form, engaging clips from long-form video content automatically. It analyzes transcripts and video data to identify the most engaging or “viral” moments, reducing the need for manual editing. The system integrates tools like FFmpeg and AI models to handle segmentation, cropping, and formatting for vertical video platforms. It supports automation workflows that allow creators to produce multiple clips efficiently at scale. The project focuses on content repurposing, helping users adapt long videos into formats suitable for platforms like TikTok and YouTube Shorts. ... -
14
dj-stripe
dj-stripe automatically syncs your Stripe Data to your local database
Dj-stripe is an extensible wrapper around the Stripe API that continuously syncs most of the Stripe Data to your local database as pre-implemented Django Models, out of the box! This allows you to use the Django ORM, in your code, to work with the data making it easier and faster! For example, if you need to interact with a customer subscription, you can use dj-stripe’s Subscription Model, in your code, to get the subscription data for that customer as well as the related models’ data too (if need be and potentially in 1 database query!) ... -
15
NGINX Admin’s Handbook
How to improve NGINX performance, security, and other important things
...The handbook spans fundamentals and advanced topics alike, from HTTP and SSL/TLS basics to reverse proxy patterns, performance tuning, debugging workflows, and hardening strategies. A centerpiece is its prioritized checklist of 79 rules, grouped by criticality, helping readers focus on what most impacts security, reliability, and speed. Instead of copy-paste snippets in isolation, it emphasizes understanding trade-offs, avoiding common pitfalls, and balancing security with usability. Designed for system administrators and web application engineers, it aims to be a living companion that encourages experimentation, measurement, and continuous improvement of NGINX configurations -
16
Letterboxd Recommendations
Scraping publicly-accessible Letterboxd data for movie recommendations
...A user's "star" ratings are scraped from their Letterboxd profile and assigned numerical ratings from 1 to 10 (accounting for half stars). Their ratings are then combined with a sample of ratings from the top 4000 most active users on the site to create a collaborative filtering recommender model using singular value decomposition (SVD). All movies in the full dataset that the user has not rated are run through the model for predicted scores and the items with the top predicted scores are returned. Due to constraints in time and computing power, the maximum sample size that a user is allowed to select is 500,000 samples, though there are over five million ratings in the full dataset from the top 4000 Letterboxd users alone. -
17
esp32-ai
Running a 28.9M parameter LLM on an $8 microcontroller
esp32-ai is an experimental language-model project that runs a 28.9-million-parameter model entirely on an ESP32-S3 microcontroller. The quantized model occupies about 14.9 MB and generates text without sending data to a server. Most parameters remain in flash through a Per-Layer Embeddings design, while active computation uses SRAM and PSRAM. This memory layout allows the device to retrieve only the embedding rows required for each token. The implementation reaches roughly 9.5 tokens per second and can display generated words on a connected screen. Trained on TinyStories, the model produces short, simple stories rather than answering questions, following instructions, or providing factual knowledge. ... -
18
MiniMind-O
A 0.1B Omni model trained from scratch
...It includes both mini and full training data paths, allowing learners to run a complete workflow quickly or reproduce the released model setup more closely. The implementation emphasizes native PyTorch code instead of relying on high-level third-party abstractions. minimind-o is most useful for developers and researchers who want to understand how multimodal and speech-capable AI systems are built from the ground up. -
19
R-KV
Redundancy-aware KV Cache Compression for Reasoning Models
...R-KV introduces a method for compressing the KV cache during decoding, allowing models to maintain reasoning performance while reducing memory consumption and computational overhead. The approach focuses on identifying which attention heads and cache components are most important for maintaining reasoning quality, allowing less critical information to be compressed or discarded. This results in more efficient inference without significantly degrading model performance. -
20
MoBA
MoBA: Mixture of Block Attention for Long-Context LLMs
...The architecture adapts ideas from Mixture-of-Experts networks and applies them directly to the attention mechanism of transformer models. Instead of forcing each token to attend to every other token in the sequence, MoBA divides the context into blocks and dynamically routes queries to only the most relevant segments of information. This routing strategy reduces the computational cost associated with traditional attention while preserving performance on reasoning and long-context tasks. The approach allows language models to scale to significantly longer input contexts without the quadratic computational cost normally associated with transformer attention mechanisms. -
21
Google CTF
Google CTF
Google CTF is the public repository that houses most of the challenges from Google’s Capture-the-Flag competitions since 2017 and the infrastructure used to run them. It’s a learning and practice archive: competitors and educators can replay tasks across categories like pwn, reversing, crypto, web, sandboxing, and forensics. The code and binaries intentionally contain vulnerabilities—by design—so users can explore exploit chains and patching in realistic settings. -
22
CLIP
CLIP, Predict the most relevant text snippet given an image
CLIP (Contrastive Language-Image Pretraining) is a neural model that links images and text in a shared embedding space, allowing zero-shot image classification, similarity search, and multimodal alignment. It was trained on large sets of (image, caption) pairs using a contrastive objective: images and their matching text are pulled together in embedding space, while mismatches are pushed apart. Once trained, you can give it any text labels and ask it to pick which label best matches a given... -
23
UForm
Multi-Modal Neural Networks for Semantic Search, based on Mid-Fusion
...Due to independent encoding late-fusion models are good at capturing coarse-grained features but often neglect fine-grained ones. This type of models is well-suited for retrieval in large collections. The most famous example of such models is CLIP by OpenAI. Early-fusion models encode both modalities jointly so they can take into account fine-grained features. Usually, these models are used for re-ranking relatively small retrieval results. Mid-fusion models are the golden midpoint between the previous two types. Mid-fusion models consist of two parts – unimodal and multimodal. -
24
pybaselines
Library of algorithms for baseline correction of experimental data
...The aim of the project is to provide a semi-unified API to allow quick testing and comparing multiple baseline correction algorithms to find the best one for a set of data. pybaselines has 50+ baseline correction algorithms. These include popular algorithms, such as AsLS, airPLS, ModPoly, and SNIP, as well as many lesser-known algorithms. Most algorithms are adapted directly from literature, although there are a few that are unique to pybaselines, such as penalized spline versions of Whittaker-smoothing-based algorithms. The full list of implemented algorithms can be found in the documentation. -
25
Shynet
Modern, privacy-friendly, and detailed web analytics
Modern, privacy-friendly, and detailed web analytics that works without cookies or JS. There are a lot of web analytics tools. Unfortunately, most of them come with the following caveats. They require handing all of your visitors' info to a third-party company They use cookies to track visitors across sessions, so you need to have those annoying cookie notices. They collect so much personal data that even the NSA is jealous. They are closed source and/or expensive, often with limited data portability. ...
