Hydra Landscape
The Hydra Landscape is a curated directory of applications, frameworks, tools, plugins, and learning resources that use Hydra in a visible and meaningful way. Listings are informational and do not imply official endorsement.
Suggest a project
Open an issue or pull request with a neutral description, a project link, and a short note showing where Hydra usage is visible.
A project may be listed when:
- It has a publicly accessible repository or project page.
- It uses, extends, integrates with, or teaches Hydra in a clear way.
- Hydra usage is visible in documentation, examples, configuration, or code.
- It is at least minimally maintained, or remains useful as an archival example.
- Its listing can be described neutrally and factually.
Featured projects are selected separately by Hydra's maintainers. They must be actively maintained, well documented, technically strong, and ready for the current Hydra release.
Use the badge
Projects configured with Hydra can add this badge to their README:
[](https://hydra.cc)
Showing 104 of 104 projects
Aequitas Flow is a package-owned Fair ML experimentation component that intentionally uses Hydra for its configuration structure. Hydra is installed through Aequitas's optional flow dependency group, so its reach is optional rather than core.
AWS recipes for training, fine-tuning, evaluating, and reinforcing foundation models on SageMaker HyperPod.
The RL Bridge for LLM-based Agent Applications. Made Simple & Flexible.
Benchmarking Agentic LLM and VLM Reasoning On Games
The independently named Benchmark experiment runner uses Hydra as its entry point and configuration-space engine for Mountpoint experiments. It also implements and activates a custom Hydra Sweeper for benchmark-specific parameter combinations.
BenchMARL is a library for benchmarking Multi-Agent Reinforcement Learning (MARL). BenchMARL allows to quickly compare different MARL algorithms, tasks, and models while being systematically grounded in its two core tenets: reproducibility and standardization.
BiomedParse: A Foundation Model for Joint Segmentation, Detection, and Recognition of Biomedical Objects Across Nine Modalities
BioNeMo Recipes: For building and adapting AI models in drug discovery at scale
A benchmark dataset collection for bird sound classification
CALVIN is a simulated benchmark for language-conditioned robotic manipulation. Its baseline-agent training subsystem directly exposes Hydra configuration overrides to users.
High-performance inference framework for large language models, focusing on efficiency, flexibility, and availability.
ClearML - Auto-Magical CI/CD to streamline your AI workload. Experiment Management, Data Management, Pipeline, Orchestration, Scheduling & Serving in one MLOps/LLMOps solution
A ComfyUI custom node designed for advanced image background removal and object, face, clothes, and fashion segmentation, utilizing multiple models including RMBG-2.0, INSPYRENET, BEN, BEN2, BiRefNet, SDMatte, SAM, SAM2, SAM3 and GroundingDINO.
First-party Hydra tool that generates Structured Config dataclasses from Python class signatures for use with instantiate().
Cross-view Transformers for real-time Map-view Semantic Segmentation (CVPR 2022 Oral)
A Cookiecutter data-science starter template whose generated project uses Hydra for configuration. The template includes executable Hydra integration and user guidance for viewing and overriding configurations.
Python Package for Airborne RGB machine learning
Deep learning quantum Monte Carlo for electrons in real space
detrex is a research platform for DETR-based object detection, segmentation, pose estimation and other visual recognition tasks.
DIAMOND (DIffusion As a Model Of eNvironment Dreams) is a reinforcement learning agent trained in a diffusion world model. NeurIPS 2024 Spotlight.
code for "Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion"
[ICML 2025] Official PyTorch Implementation of "History-Guided Video Diffusion"
An I/O benchmark for deep Learning applications
🧬 Generative modeling of regulatory DNA sequences with diffusion probabilistic models 💨
Open-source deep-learning framework for exploring, building and deploying AI weather/climate workflows.
Ego4D Features is a documented feature-extraction component of the Ego4d Python module. It natively uses Hydra for its configurable feature-extraction workflows.
EgoVerse’s egomimic component is the repository-owned behavior-cloning and flow-matching training application. Its main training entry point directly initializes Hydra and composes repository-owned training and Submitit launcher configuration.
ESPnet3 intentionally uses Hydra for configuration-driven object construction in production training and trainer paths. The dependency is explicitly scoped to ESPnet3 and accompanied by project-specific configuration documentation.
Espresso: A Fast End-to-End Neural Speech Recognition Toolkit
ExecuTorch’s LLM Export API directly uses Hydra to provide structured YAML and command-line configuration for exporting models. The Hydra integration is maintained, repository-native subsystem code rather than a copied example, vendored component, or transitive dependency.
Official code for "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"
FAIRChem’s fairchem-core component natively uses Hydra for its primary training/job configuration and runner instantiation. The project provides machine-learning methods for chemistry and materials science.
A Fast Deep Learning Model to Upsample Low Resolution Videos to High Resolution at 30fps
[CVPR 2025] Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass
FlagScale is a large model toolkit based on open-sourced projects.
[3DV 2025] Code for "FlowMap: High-Quality Camera Poses, Intrinsics, and Depth via Gradient Descent" by Cameron Smith*, David Charatan*, Ayush Tewari, and Vincent Sitzmann
flyteplugins-hydra is an optional Flyte SDK plugin that provides a Hydra launcher plus CLI and Python entry points for running Hydra-composed Flyte tasks and sweeps. Its implementation is native integration code rather than an example or copied tutorial.
Central repository for biomolecular foundation models with shared trainers and pipeline components
FusionBench: A Comprehensive Benchmark/Toolkit of Deep Model Fusion
End-to-end SFM framework based on GTSAM
Habitat-Lab is a modular embodied-AI library whose own configuration system migrated to Hydra. Its primary configuration API registers a project-defined Hydra search-path plugin.
Hydra launcher plugin that runs jobs on Modal, with config-driven sandbox customisation.
Create powerful Hydra applications without the yaml files and boilerplate code.
Official implementation for HyenaDNA, a long-range genomic foundation model built with Hyena
Imitation learning algorithms
Transformers are Sample-Efficient World Models. ICLR 2023, notable top 5%.
GCM Physics written in JAX
Large Concept Models: Language modeling in a sentence representation space
PyTorch Lightning + Hydra. A very user-friendly template for ML experimentation. ⚡🔥⚡
MapAnything: Universal Feed-Forward Metric 3D Reconstruction
An AI framework for generating and modding osu! beatmaps for all gamemodes from spectrogram inputs.
Official code repo for the MARL book (www.marl-book.com)
[ICASSP 2024] 🍵 Matcha-TTS: A fast TTS architecture with conditional flow matching
Official implementation of MatterGen -- a generative model for inorganic materials design across the periodic table that can be fine-tuned to steer the generation towards a wide range of property constraints.
🦁 A research-friendly codebase for fast experimentation of multi-agent reinforcement learning in JAX
McByte is an independently named, clean-room-adapted tracker subsystem within the Trackers library. Its optional SAM/Cutie mask pipeline directly uses Hydra to initialize and compose Cutie configurations, while the default tracker configuration leaves mask management disabled.
[NeurIPS'23 Spotlight] "Mind2Web: Towards a Generalist Agent for the Web" -- the first LLM-based web agent and benchmark for generalist web agents
MyoSuite’s maintained agent-baseline subsystem directly uses Hydra as the configuration and launch interface for MJRL and Stable-Baselines3 training. Hydra is not required by the simulator’s core package.
NequIP is a code for building E(3)-equivariant interatomic potentials
A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)
open-metric-learning is a metric-learning library whose optional Pipelines component uses Hydra for configuration-driven experiment entry points. The repository documents Hydra as its config parser and provides pipeline examples with Hydra-decorated launch functions.
[NeurIPS D&B '25] The one-stop repository for LLM unlearning
🏋️ A unified multi-backend utility for benchmarking Transformers, Timm, PEFT, Diffusers and Sentence-Transformers with full support of Optimum's hardware optimizations & quantization schemes.
Source Code for Paper "OrienterNet Visual Localization in 2D Public Maps with Neural Matching"
Agent-R1’s Paper Search Recipe owns a Hydra configuration for its offline inference and evaluation path. The recipe is a maintained component of the active Agent-R1 agentic reinforcement-learning framework.
PARTNR Planner is a Habitat-based embodied multi-agent planning application. Its project-owned habitat_llm evaluation entry point is a Hydra application backed by a repository-wide configuration tree.
PDEBench natively uses Hydra as the configuration entry point for its forward baseline-model training subsystem. This is a maintained scientific-machine-learning benchmark and tooling project.
PhysicsNeMo’s Lagrangian MeshGraphNet reference sample directly runs its training program through Hydra. The usage is an NVIDIA-authored, documented training-recipe capability, while the repository describes training recipes as separate from packaged core artifacts.
PINNs-Torch, Physics-informed Neural Networks (PINNs) implemented in PyTorch.
Point Transformers
Comprehensive benchmarking of protein-ligand structure prediction methods. (Nature Machine Intelligence)
[CVPR 2024] Probing the 3D Awareness of Visual Foundation Models
Quadra: Effortless and reproducible deep learning workflows with configuration files.
Hydra use belongs to the optional DeepSpeech2 model integration, not to the AI Hub Models core. The integration adapts the upstream DeepSpeech2 implementation for Qualcomm-device export and explicitly installs its Hydra-related configuration dependency.
A PyTorch library for all things Reinforcement Learning (RL) for Combinatorial Optimization (CO)
Democratizing Reinforcement Learning for LLMs
🤖 The Full Process Python Package for Robot Learning from Demonstration and Robot Manipulation
Structured state space sequence models
SAM 3D Objects
SAM-PT: Extending SAM to zero-shot video segmentation with point-based tracking.
[CoRL 2023] This repository contains data generation and training code for Scaling Up & Distilling Down
SchNetPack - Deep Neural Networks for Atomistic Systems
The independently named siclib training library uses Hydra to compose and override GeoCalib training configurations. Hydra is a training/evaluation subsystem rather than a dependency of the lightweight inference package.
Strands-based agents and harnesses for agentic benchmarks.
SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning
A general physic-based retargeting framework.
Official implementation of “Splatter Image: Ultra-Fast Single-View 3D Reconstruction” (CVPR 2024).
🏛️A research-friendly codebase for fast experimentation of single-agent reinforcement learning in JAX • End-to-End JAX RL
Monty is a sensorimotor learning framework based on the thousand brains theory of the neocortex.
This is the official repository for the paper TLOB: A Novel Transformer Model with Dual Attention for Price Trend Prediction with Limit Order Book Data.
TopoBench is a Python library designed to standardize benchmarking and accelerate research in Topological Deep Learning
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.
tsl: a PyTorch library for processing spatiotemporal data.
[NeurIPS 2025] TTRL: Test-Time Reinforcement Learning
Official implementation of Tuna-2: Pixel Embeddings Beat Vision Encoders for Unified Understanding and Generation
Unified Training of Universal Time Series Forecasting Transformers
UniRL is a Framework for Unified Multimodal Model Reinforcement Learning
verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework
verl-agent is an extension of veRL, designed for training LLM/VLM agents via RL. verl-agent is also the official code for paper "Group-in-Group Policy Optimization for LLM Agent Training"
Modular imitation-learning framework for training, compressing, and deploying robot manipulation policies.
VGGSfM: Visual Geometry Grounded Deep Structure From Motion
[ICLR '26 - Virne] A simulator & benchmark for resource allocation (RA) problems in network function virtualization (NFV), i.e., NFV-RA, including virtual network embedding, service function chain deployment, network slicing, etc.
The AI developer platform. Use Weights & Biases to train and fine-tune models, and manage models from experimentation to production.
[NeurIPS 2025] WorldMem: Long-term Consistent World Simulation with Memory
How projects are categorized
Good Hydra usage highlights projects that make thoughtful use of Hydra's configuration model and provide a strong experience for their users.
Powered by Hydra includes projects with meaningful Hydra usage without implying that their configuration architecture is an example to follow.
Projects may also extend, integrate with, or teach Hydra. Tags describe relevant topics and technologies; they are not intended to form a rigid domain hierarchy.