GPU-ready training and inference images with drivers, CUDA and frameworks already matched to each other. Browse all 247 products in this category from WIEWAVE's published marketplace catalog — each one a hardened, maintained image you can deploy from your own cloud account.
Search the full catalogSelect any image for its supported distributions, marketplace availability and common questions.
A development environment preconfigured with LangChain and LangFlow for building and visually prototyping LLM-powered applications.
An open-source toolkit with metrics and algorithms for detecting and mitigating bias in machine learning models.
A machine learning application that adjusts product prices in real time based on demand, competition, and other signals.
A development environment bundling Python, Jupyter notebooks, and common machine learning libraries for data science work.
A Python library providing fast, GPU-friendly image augmentation transforms for training computer vision models.
An open-source natural language processing research library built on PyTorch, developed by the Allen Institute for AI.
A full-stack application for building private, document-aware chatbots on top of local or hosted large language models.
Automated machine learning toolkit built on scikit-learn that searches pipelines and hyperparameters using Bayesian optimisation, shipped with Python and its numeric dependencies.
Lightweight Python framework for building LLM agents with custom tools, including simulated conversations that let you evaluate agent behaviour automatically.
AutoML library from AWS that trains and ensembles models for tabular, text, image and time-series data with only a few lines of Python.
Experimental autonomous agent platform chaining LLM calls, tools and memory to pursue user-defined goals, preinstalled with its Python runtime and web interface.
AutoML system built on Keras and TensorFlow that searches neural network architectures and hyperparameters for classification, regression and image tasks.
Curated collection of AI tools and references packaged as a browsable image; confirm the exact bundled contents with the publisher before production use.
Task-driven autonomous agent that combines an LLM with a vector store to generate, prioritize and execute task lists toward a stated objective.
Python framework for packaging machine learning models into production services with REST and gRPC endpoints, adaptive batching, containerisation and a model store.
Transformer language model for text classification, question answering and embeddings, shipped with pretrained weights, tokenizers and a framework for fine-tuning.
Class-conditional generative adversarial network for high-resolution image synthesis, provided with pretrained weights and inference scripts for sampling from ImageNet categories.
Text-to-image generation tool that steers BigGAN with CLIP guidance to synthesise pictures from prompts, installed with PyTorch and GPU dependencies.
Multilingual large language model fine-tuned from BLOOM for instruction following across dozens of languages, served with transformers and inference dependencies.
Deep learning framework from Berkeley Vision and Learning Center, shipped with Python bindings, the reference model zoo and CPU inference dependencies.
French-language RoBERTa model packaged with Hugging Face Transformers and PyTorch for text classification, named entity recognition and other NLP tasks.
Gradient boosting library from Yandex with native categorical feature handling, installed alongside Python bindings, GPU training support and Jupyter notebooks.
Define-by-run deep learning framework for Python, preinstalled with CuPy GPU arrays, NumPy and Jupyter for building and training neural networks.
Python library for building conversational bots that learn from example dialogue corpora, installed with its storage adapters and training utilities.
Preconfigured MLOps stack combining pipeline orchestration, a model registry and automated training, evaluation and deployment steps for continuous delivery of models.
Development environment with Anthropic's Claude command-line tooling and API SDKs installed for building applications and agents against Claude models.
MLOps platform bundling experiment tracking, dataset versioning, pipeline orchestration and remote agents, backed by MongoDB, Elasticsearch and Redis on the server.
Meta's code-specialised Llama model served locally for completion, infilling and generation, with GPU inference runtime and dependencies preinstalled.
Seven-billion-parameter Code Llama checkpoint packaged for single-GPU inference, suited to code completion and translation between programming languages.
Instruction-tuned 7B Code Llama variant served through Ollama, answering natural-language prompts about code behind an OpenAI-compatible local API endpoint.
Sourcegraph's AI coding assistant configured to index a repository and answer questions, explain code and generate patches using codebase context.
Structure-from-motion and multi-view stereo pipeline that reconstructs 3D geometry and camera poses from image collections, built with CUDA acceleration.
Distributed training system for large models offering tensor, pipeline and sequence parallelism plus ZeRO memory optimisation across multi-GPU nodes.
Natural language processing image for extracting entities, sentiment and key phrases from text; the specific engine and models depend on the publisher's build.
Command-line environment intended for image analysis tasks such as detection and classification; the bundled vision libraries depend on the publisher's build.
Machine learning image for sorting text or media into predefined labels; the underlying models and framework depend on the publisher's configuration.
Conditioning framework for Stable Diffusion that steers image generation using pose, depth, edge and segmentation maps alongside the text prompt.
Text-to-speech toolkit shipping pretrained models for speech synthesis and voice cloning, plus training scripts for building custom voices.
Scene text detection model that locates character regions and links them into word boxes, typically used as an OCR preprocessing stage.
Python framework for orchestrating multi-agent LLM workflows, assigning roles, tasks and tools to agents that collaborate toward a shared goal.
Image for hosting or fine-tuning self-selected large language models; the served models and inference stack depend on the publisher's configuration.
Conda, NumPy, pandas, scikit-learn and JupyterLab as one coherent environment, with BLAS linked against a threaded build.
Cascaded pixel-space diffusion model for text-to-image generation that renders legible text inside images, packaged with GPU inference dependencies.
Markerless pose estimation toolbox that tracks animal and human body parts in video using transfer learning on deep neural networks.
Semantic segmentation architecture using atrous convolution and spatial pyramid pooling, supplied with pretrained weights for per-pixel image labeling.
Conversational AI library for building chatbots and NLP pipelines, with pretrained models for intent classification, named entity recognition and question answering.
Inference image that serves DeepSeek and Llama open-weight language models on GPU hardware for chat, coding and reasoning workloads.
Add-in that brings DeepSeek language model responses into spreadsheet cells, letting formulas generate, summarise, translate and classify worksheet content.
Code-oriented large language model of 33 billion parameters, instruction tuned for completion, explanation and repository-level tasks and served for GPU inference.
Reasoning-focused language model served behind a browser chat interface, giving a private endpoint for step-by-step prompting on GPU instances.
Python library built on PyTorch for training and running object detection models over images and video through a small high-level API.
Dify provides a low-code platform for building LLM applications, bundling workflow orchestration, RAG pipelines, agents and prompt management over PostgreSQL and Redis.
Compact transformer language model distilled from BERT, shipped with Hugging Face Transformers and PyTorch for text classification and embedding workloads.
dlib is a C++ toolkit with Python bindings for machine learning, face detection and landmark estimation, compiled with its image processing modules.
Python library that parses PDFs, DOCX and presentations into structured Markdown or JSON for retrieval pipelines, installed with its layout and table models.
Runs Docling behind an HTTP API, exposing document conversion endpoints and an OpenAPI schema so other services can submit files for parsing.
Fine-tuning setup for diffusion models that teaches a subject from a handful of images, with PyTorch, CUDA and training scripts preinstalled.
Optical character recognition library covering more than eighty languages, packaged with PyTorch, detection and recognition models, and a short Python API.
Python library for readable tensor reshaping, reduction and repetition across NumPy, PyTorch and TensorFlow, installed in a prepared data science environment.
Transformer model pretrained with replaced-token detection, provided with Hugging Face Transformers for fine-tuning on classification and question answering tasks.
AI & Machine LearningEnvironment for EleutherAI open models such as GPT-Neo, GPT-J and Pythia, with Transformers, PyTorch and CUDA libraries ready for inference.
Python package that explains machine learning classifiers by showing feature weights and per-prediction breakdowns for scikit-learn, XGBoost and text pipelines.
Python environment prepared for generating vector embeddings from text with common open-source models; the specific toolkit behind this listing is unspecified.
Text processing environment for extracting entities such as people, organisations and locations; the underlying NLP library for this listing is unspecified.
Generative adversarial network for image super-resolution, shipped with pretrained upscaling weights and PyTorch inference scripts for enhancing low-resolution photos.
ExLlama is a fast GPU inference library for running quantized LLaMA-family large language models, with V2 and V3 supporting newer quantization formats.
Explainable AI CLI appears to be a command-line tool for generating model interpretability reports, though the specific vendor implementation is unconfirmed.
FaceSwap is a deep-learning tool for creating face-swapped images and video using neural network models trained on GPU hardware.
FastAI is a deep learning library built on PyTorch that provides high-level components for training neural networks with fewer lines of code.
FastDeploy is a toolkit for deploying and serving trained deep learning models across CPUs and GPUs with optimised inference backends.
FastNLP is a Python natural language processing toolkit providing modular components for building and training NLP models.
fastText is a library from Facebook AI Research for efficient learning of word embeddings and fast text classification.
Flair is a Python natural language processing framework built on PyTorch that provides pretrained models for named entity recognition and text classification.
FLAML is a lightweight Python AutoML library from Microsoft that automatically selects and tunes machine learning models with low computational cost.
Flan-T5 is an instruction-tuned variant of Google's T5 language model, fine-tuned on a wide range of tasks for improved zero-shot performance.
Flowise is a drag-and-drop, low-code tool for visually building large language model applications and AI agent workflows.
Flowise AI is a visual, low-code builder for chaining large language models, tools and data sources into custom AI workflows.
Flowtron is NVIDIA's flow-based generative neural network for synthesising expressive and controllable text-to-speech audio.
Fooocus is a simplified Stable Diffusion image generation interface, and this image adds an API server for programmatic image generation requests.
Research toolkit that finds interpretable latent directions in pretrained GANs using PCA, letting you edit generated images along semantic axes.
Gencore AI connector that ingests documents from Amazon S3 buckets into governed retrieval pipelines, preserving source permissions for enterprise LLM applications.
Builds retrieval pipelines over Azure Blob Storage containers, syncing unstructured files into Gencore AI with entitlement-aware indexing for internal AI assistants.
Connector image reading unstructured content from Azure Data Lake Storage Gen2 hierarchies and preparing it as vectorised, access-controlled knowledge for Gencore AI.
Ingests files from Box workspaces into Gencore AI, keeping folder-level sharing permissions attached so generated answers respect existing collaboration boundaries.
Pulls documents from Dropbox into Gencore AI pipelines and indexes them for retrieval-augmented chat while carrying over per-file access rights.
Connects Google Cloud Storage buckets to Gencore AI so object contents are chunked, embedded and served to LLM applications under governance controls.
Reads OneDrive for Business libraries into Gencore AI, syncing personal and shared files into permission-aware vector indexes for enterprise assistants.
Mounts NFS exports and feeds their unstructured files into Gencore AI, producing curated knowledge pipelines for retrieval-augmented generation over network shares.
Ingestion connector for Oracle Cloud Infrastructure Object Storage that channels bucket contents into the Gencore AI indexing, entitlement and safety pipeline.
Extracts attachments and knowledge articles from ServiceNow into Gencore AI so support copilots answer from ticket data under existing access rules.
Collects files from remote SFTP servers on a schedule and loads them into Gencore AI for embedding, indexing and governed LLM retrieval.
Streams Slack Enterprise Grid conversations and shared files into Gencore AI, building searchable, permission-aware context without exposing private channels to everyone.
Indexes Windows SMB and CIFS file shares through Gencore AI, mapping share permissions onto retrieval so answers use only documents a user may read.
Python library for topic modelling and document similarity providing Word2Vec, Doc2Vec, LDA and streaming corpora that exceed available memory.
Extends pandas with geometry columns so spatial joins, reprojection and shapefile or PostGIS input work through a familiar dataframe API.
Apache MXNet with the Gluon imperative API for defining and training neural networks, including pretrained vision and language model zoos.
Development image preloaded with the Google Cloud AI Platform (Vertex AI) Python SDK and gcloud CLI for training, tuning and deploying models.
Base compute image with NVIDIA drivers, CUDA toolkit and container runtime configured so GPU workloads such as training or inference run without further driver setup.
Data science environment combining the H2O machine learning library with NumPy and pandas plus Jupyter, ready for tabular modelling workflows.
Open source H2O platform for distributed machine learning, exposing AutoML, GLM, GBM and deep learning through Flow, Python and R interfaces.
Open source framework from deepset for building retrieval-augmented generation and search pipelines, connecting document stores, embedders, retrievers and LLM providers.
Environment preloaded with Transformers, Datasets and Tokenizers plus PyTorch, ready for fine-tuning and serving natural language models pulled from the Hub.
Text Generation Inference with tensor parallelism and continuous batching configured for GPU serving at real concurrency.
Platform for natural language data work, letting teams cluster, label and curate conversational utterances into intents and training sets for NLU models.
Python library for hyperparameter search using random, TPE and adaptive algorithms, installed alongside SciPy and a Jupyter environment for distributed trials.
Tooling and SDK access for Google's Imagen text-to-image models, covering image generation, editing, upscaling and captioning through Vertex AI endpoints.
AI & Machine LearningMicrosoft framework for model-based machine learning that lets .NET developers describe probabilistic graphical models and run Bayesian inference over them.
Cross-lingual pretrained language model from Microsoft trained with information-theoretic objectives, used for multilingual classification and retrieval after fine-tuning.
Toolkit for optimising and running deep learning inference on Intel CPUs and integrated GPUs, including model conversion, quantisation and the runtime.
Stable Diffusion interface with a node-based canvas, inpainting, ControlNet support and model management, served as a web application on GPU instances.
NumPy-compatible Python library for autodiff and XLA-accelerated array computation, installed with the runtime needed for CPU or GPU numerical workloads.
Python framework for building multimodal and neural search services, exposing gRPC and HTTP endpoints for embedding pipelines and vector retrieval.
Just-in-time compiled deep learning framework using meta-operators and unified graph execution, installed with Python bindings for model training.
Python utilities for lightweight pipelining: transparent disk caching of function results, parallel loops and efficient persistence of large NumPy arrays.
Minimal Jupyter base environment from the official docker-stacks lineage, giving a conda-managed Python kernel to build custom notebook images on.
Interactive notebook server for Python, letting you mix code, plots and prose in the browser; includes the IPython kernel and scientific packages.
JupyterHub with per-user notebook servers, an authenticator configured and the scientific Python stack pre-installed.
Jupyter's tabbed workspace interface with file browser, terminals, notebooks and extension support, served over HTTP with token authentication.
Python framework for structuring reproducible data science pipelines, providing a data catalog, node abstractions and pipeline visualisation tooling.
Deep learning stack pairing the Keras high-level API with a TensorFlow backend, preinstalled with Python bindings for training and inference.
Single-binary inference server for GGUF language models built on llama.cpp, exposing a chat web UI and an OpenAI-compatible API.
Differentiable computer vision library for PyTorch, supplying image transforms, geometry, filtering and feature operators as GPU-ready tensor operations.
Kubernetes custom resources for serving machine learning models, handling autoscaling, canary rollouts and standardised inference endpoints across frameworks.
Preinstalled LangChain libraries alongside Flowise, a drag-and-drop builder for wiring LLM chains, agents and retrieval flows through a browser canvas.
Combines the LangChain framework with Langflow's visual editor so you can prototype retrieval-augmented generation pipelines and export them as runnable Python code.
Visual low-code environment for composing LLM applications, running the Langflow server with its Python backend, component library and web canvas.
CompVis latent diffusion research code for text-to-image and image synthesis, with PyTorch, checkpoint loading scripts and GPU inference dependencies installed.
Environment for building conversational bot interfaces, carrying the SDKs and runtime dependencies needed to develop, test and host text and voice dialogs.
Self-hosted chat front end supporting multiple model providers, with MongoDB storage, user authentication and conversation history served by a Node.js backend.
AutoML framework that builds tabular models end to end, covering feature preprocessing, hyperparameter search and model blending from a Python API.
Gradient boosting framework tuned for large tabular datasets, installed with Python bindings, OpenMP support and the command-line training tool.
Meta's seven-billion-parameter Llama 2 model with tokenizer and PyTorch inference scripts prepared for GPU-based text generation and fine-tuning.
Deployment of Meta's 70-billion-parameter Llama 3 model, sized for multi-GPU inference with transformer libraries and CUDA dependencies preinstalled.
Eight-billion-parameter Llama 3 chat model served through a Python inference stack, suited to single-GPU generation and lightweight fine-tuning experiments.
Compiled llama.cpp HTTP server exposing an OpenAI-compatible API for running quantized GGUF models on CPU or GPU hardware.
Data framework connecting LLMs to private documents, providing indexing, retrieval and query engines in Python with common loader and vector store integrations.
Tooling image aimed at inspecting and evaluating large language model behaviour; the packaged application's exact feature set should be confirmed before use.
Drop-in OpenAI-compatible API server that runs local models for text, embeddings, images and audio without calling external inference services.
Transformer architecture using sliding-window attention to process long documents, supplied with pretrained checkpoints and Hugging Face inference dependencies.
Declarative machine learning framework where models are defined in YAML configuration, wrapping PyTorch for training text, tabular and image tasks.
Fast C++ toolkit for training and serving neural machine translation models, compiled with CPU and GPU decoding support and example configurations.
Image providing a large language model runtime for text generation workloads, with GPU inference dependencies installed and a serving entry point.
Python plotting library for static charts and figures, installed with NumPy and headless rendering backends so images can be generated server-side.
Applied reinforcement learning framework for building and training agents, offering configurable environments, distributed rollouts and PyTorch-based policy implementations.
Google's cross-platform perception framework, shipped with Python bindings and pretrained pipelines for face, hand, pose and object detection on images or video.
Speech-to-text transcription service from Meetrix that accepts audio uploads over a REST API and returns written transcripts of meetings and recordings.
Deep learning framework from Megvii with unified training and inference paths, installed as a Python package with CPU and GPU execution support.
Tooling image related to the Midjourney text-to-image generation service; the service itself is proprietary and hosted, so packaged components vary.
Huawei's deep learning framework with automatic differentiation and graph-mode execution, installed with Python bindings for model training and inference.
Packages Mistral 7B weights together with an inference server and Python dependencies so several fine-tuned variants can be served from one GPU instance.
Ships both Mistral 7B and the long-context MistralLite 7B checkpoints with a serving stack, letting either model be loaded for text generation.
MLflow tracking server with a PostgreSQL backend and object-store artifact root, so experiments survive the instance.
C++ machine learning library covering classification, regression, clustering and dimensionality reduction, installed with headers, command-line tools and Python bindings.
OpenMMLab toolbox for text detection and recognition, providing PyTorch models and inference scripts for OCR on documents and scene images.
Training and fine-tuning codebase for large language models from MosaicML, shipping Composer, MPT recipes and GPU-ready PyTorch dependencies.
Multi-dimensional image viewer for Python used in microscopy and scientific imaging, with layered visualisation, annotation tools and a plugin ecosystem.
Shell environment preloaded with Python natural language processing libraries and models for tokenization, classification, entity extraction and other text tasks.
Python framework for building and simulating large-scale spiking neural network models of the brain, installed with its solvers and visualisation interface.
Experiment tracking and model registry client for machine learning teams, logging metrics, parameters, artifacts and hardware usage from training runs.
A viewer application for inspecting the architecture, layers and weights of neural network and machine learning model files.
An NVIDIA research framework that reconstructs detailed 3D surface models from RGB video using neural rendering techniques.
An open-source TensorFlow-based deep learning platform for building and deploying convolutional neural networks for medical image analysis.
A Python library providing tools and corpora for natural language processing tasks such as tokenization, tagging and parsing.
An NVIDIA application framework and SDK collection for building and deploying AI models in medical imaging and healthcare workflows.
GPU driver and CUDA toolkit versions pinned to a combination that's actually been tested together, with nvidia-smi verified at build time.
Ollama for running open-weight models locally, with GPU offload configured and the API bound for use behind a proxy.
A bundle pairing the Ollama local large language model runtime with the Open WebUI chat interface for running open-source models.
An open format and runtime for representing and executing machine learning models across different frameworks and hardware backends.
Reference PyTorch implementation for music source separation, splitting audio tracks into vocals, drums, bass and residual stems using pretrained models.
An open-source platform for building and running autonomous AI agents that can use tools and complete multi-step tasks.
A pretrained transformer-based language model released by OpenAI, used for text generation and natural language research.
A bundle of the OpenAI Gym reinforcement learning environment toolkit together with the NumPy and Pandas data libraries.
Gymnasium, the actively maintained fork of OpenAI Gym, provides a standard API and set of environments for reinforcement learning research.
An open-source large language model fine-tuned for conversational use, distributed as weights for local or self-hosted inference.
A bundle of the OpenCV, Pillow and MediaPipe libraries preinstalled for computer vision, image processing and pose or face detection tasks.
Hyperparameter optimisation framework for Python featuring samplers, pruning and a study dashboard, installed alongside common machine learning libraries for tuning experiments.
Neural machine translation models from the Helsinki-NLP OPUS project, served through Marian and Transformers runtimes for translating text across many language pairs.
Baidu's open-source deep learning framework, installed with its Python API and model libraries for training and serving vision, OCR and NLP models.
Python data analysis library providing DataFrame structures for cleaning, joining and aggregating tabular data, installed with NumPy and a Jupyter environment.
Combined Python environment pairing pandas for tabular preprocessing with Hugging Face Transformers for loading, fine-tuning and running pretrained NLP and vision models.
Python stack combining pandas dataframes with LangChain so LLM chains and agents can query, summarise and reason over tabular datasets.
Python helper running SQLite queries directly against pandas DataFrames, letting analysts mix familiar SQL syntax into notebook and script workflows.
Text rewriting web tool image whose upstream project is unclear; confirm the exact features and licensing with the vendor before deploying it.
Python research framework from Meta AI for training and evaluating dialogue models, bundling task datasets, agent implementations and evaluation loops.
Python API from the Farama Foundation for multi-agent reinforcement learning, shipping classic, Atari and particle environment families alongside training dependencies.
Microsoft's compact 2.7-billion-parameter language model, packaged with Transformers and GPU drivers for local inference, evaluation and fine-tuning experiments.
Pretrained BERT-style language model for Vietnamese from VinAI, served through Transformers for classification, named entity recognition and other NLP tasks.
PIFuHD reconstructs high-resolution 3D clothed human meshes from a single RGB photo, shipping pretrained weights, PyTorch and inference scripts.
Conditional GAN implementation for paired image-to-image translation, packaged with PyTorch, training scripts and CUDA dependencies for tasks such as sketch-to-photo.
Open-source code generation language model trained on twelve programming languages, supplied with weights and a Hugging Face inference environment.
Polyglot is a Python natural language processing library offering tokenization, language detection, named entity recognition and word embeddings across many languages.
Microsoft toolkit for building LLM applications as executable DAGs, supporting prompt variants, batch evaluation runs, tracing and a visual editor.
Python library that generates structured prompts for NLP tasks such as classification and named entity recognition, returning parsed output from language models.
Forecasting library from Meta that fits additive time-series models with trend, seasonality and holiday effects, callable from Python and R.
Modular Python machine learning library covering neural networks, reinforcement learning and unsupervised algorithms; the project is no longer actively maintained.
Low-code Python machine learning library automating preprocessing, model comparison, tuning and deployment across classification, regression, clustering and time-series tasks.
Probabilistic programming library for Bayesian modelling in Python, running MCMC and variational inference on a PyTensor computational backend.
Earlier PyMC release built on Theano, providing Bayesian statistical models, NUTS sampling and variational inference for Python data analysis.
Python package that extracts shape, intensity and texture radiomics features from medical images and segmentation masks for quantitative imaging research.
Python environment wrapping DeepMind's StarCraft II Learning Environment, exposing observations and actions for training reinforcement learning agents against the game API.
Jupyter Notebook and JupyterLab served on a Python stack with NumPy, pandas, scikit-learn and related libraries for interactive model development.
PyTorch with a CUDA build matched to the installed driver, so training starts on the GPU instead of silently falling back to CPU.
Python framework for computational imaging and inverse problems, providing composable linear operators and proximal optimization algorithms that run on CPU or GPU.
Preconfigured environment for packaging trained models into HTTP inference endpoints, bundling common Python serving frameworks and their dependencies for quick deployment.
Ray with the head and worker roles scripted, for distributing training, tuning and batch inference across a cluster.
Meta's applied reinforcement learning platform, formerly Horizon, built on PyTorch for training and evaluating decision policies from logged production data.
Stack for observing deployed machine learning models in production, collecting prediction data to track drift, data quality and accuracy over time.
Development image with AWS SDK and CLI tooling configured for building image and video analysis workflows against the Amazon Rekognition service.
Reinforcement learning library built on Ray that provides distributed implementations of PPO, DQN and other algorithms behind a common training API.
Transformer language model from the BERT family, packaged with Hugging Face tooling for fine-tuning on classification, extraction and other NLP tasks.
OpenAI's Bullet-based robot simulation environments for reinforcement learning research, providing locomotion and control tasks behind a Gym-compatible interface.
Environment for SAM, a name shared by several products, most commonly Meta's Segment Anything image segmentation model; verify the intended variant before use.
Machine learning library for Python covering classification, regression, clustering and model selection, installed alongside NumPy, SciPy, pandas and Jupyter.
Sequential model-based optimization library for Python that performs Bayesian hyperparameter search, built on top of scikit-learn, NumPy and SciPy.
Adds numerical routines for integration, optimisation, linear algebra, signal processing and statistics on top of NumPy in a prepared Python environment.
Statistical plotting library layered over Matplotlib, shipped with pandas and Jupyter so heatmaps, distribution and regression charts can be produced immediately.
Python workspace preloaded with NLP libraries and pretrained models for classifying text polarity, including tokenization, feature extraction and evaluation notebooks.
Collection of machine learning algorithms implemented in C++ with bindings for Python, R and Octave, focused on kernel methods and support vector machines.
Wrapper exposing PyTorch models through the scikit-learn estimator API, so grid search, pipelines and callbacks work directly with neural networks.
Data-centric AI platform where labelling functions and weak supervision generate training data programmatically instead of relying on manual annotation.
Industrial natural language processing library for Python offering tokenization, tagging, named entity recognition and pretrained pipelines for dozens of languages.
Reinforcement learning library built on PyTorch implementing PPO, DQN, SAC and A2C behind a consistent training, evaluation and logging API.
Python package for estimating statistical models, running hypothesis tests and exploring data, preinstalled with its NumPy, SciPy and pandas dependencies.
Python framework that turns plain scripts into interactive data and machine-learning web apps, with the runtime and server preconfigured on port 8501.
Neural text-to-speech architecture that generates mel spectrograms from text for a vocoder such as WaveGlow, packaged with PyTorch and CUDA dependencies.
TensorFlow with GPU support, the matching cuDNN and TensorBoard ready to serve on first launch.
Optical character recognition engine that extracts text from images and scanned documents, shipping with language training data and the Leptonica imaging library.
Gradio web interface for running and chatting with local large language models, supporting Transformers, llama.cpp and ExLlama backends plus LoRA loading.
Toolkit for converting text into dense vector embeddings and scoring semantic similarity, with pretrained sentence models in a ready Python environment.
Python framework for adversarial attacks, data augmentation and robustness testing of NLP models, integrating with Hugging Face Transformers and datasets.
Simple Python API for common natural language processing tasks including sentiment analysis, noun phrase extraction and part-of-speech tagging, built over NLTK.
Python library for defining and evaluating multi-dimensional array expressions with GPU acceleration, historically used as the backend for early deep learning frameworks.
Compact deep learning framework with autograd and several accelerator backends, aimed at readable model code and experimentation, installed in a Python environment.
Lua-based scientific computing framework with tensor math and neural network packages, the predecessor of PyTorch, built with LuaJIT and CUDA bindings.
Model serving framework for PyTorch exposing inference and management REST endpoints, with batching, versioning and metrics for deployed model archives.
NVIDIA Triton serving TensorRT, ONNX and PyTorch models side by side, with dynamic batching and the metrics endpoint enabled.
Toolkit for training reinforcement-learning agents inside Unity environments, bundling the Python trainer package, PyTorch and the ml-agents communication library.
Vectice provides model documentation and governance for data science teams, capturing experiment lineage, dataset versions and validation evidence for review.
Offline speech recognition toolkit built on Kaldi, providing streaming APIs, Python and server bindings, and compact acoustic models for many languages.
AI & Machine LearningVisual Object Tagging Tool from Microsoft for labelling images and video frames, exporting annotations in formats consumed by common object detection trainers.
Fast online machine learning library supporting contextual bandits, reinforcement learning and large-scale linear models trained from streaming input files.
Convolutional neural network upscaler that enlarges and denoises anime-style artwork and photographs, with command line and web interfaces for batch jobs.
Self-supervised speech representation model from Meta for automatic speech recognition, packaged with PyTorch, fairseq or Transformers and pretrained checkpoints.
Gradient boosting library for regression, classification and ranking tasks, preinstalled with Python bindings and the scientific stack needed to train tree ensembles.
Pretrained autoregressive transformer model for NLP tasks such as classification and question answering, packaged with PyTorch and the Transformers library.
Open-source toolkit supplied as a preconfigured image; consult the upstream project documentation for the exact functionality, dependencies and usage of this build.
MLOps framework for writing portable machine learning pipelines and switching between orchestrators, artifact stores and model registries through one Python API.
Name the product and the distribution — we'll build, harden, certify and publish it on Azure, AWS or Google Cloud Marketplace.
