Top Builders

Explore the top contributors showcasing the highest number of app submissions within our community.

Hugging Face Hub

The Hugging Face Hub is an open-source repository platform that hosts over one million machine learning models, datasets, and interactive applications. It serves as the central collaboration layer for the ML community, enabling developers to discover, share, version, and deploy models across every modality including text, vision, audio, and multimodal. Model checkpoints on the Hub are compatible with the Transformers, Diffusers, and Datasets libraries, and can be loaded in a few lines of code.

General
AuthorHugging Face
TypeML Model Repository and Collaboration Platform
Websitehuggingface.co
DocumentationHub Documentation
Repositorygithub.com/huggingface/huggingface_hub
Modelshuggingface.co/models
Datasetshuggingface.co/datasets

Start building with Hugging Face Hub

The Hub is the fastest way to get a pretrained model running in your project. Load any of the 1M+ checkpoints directly into Transformers or Diffusers with a single function call, or browse the Hub to find the right base model for your use case. You can host your own models privately and share fine-tuned adapters with the community without uploading full model weights. During AMD-sponsored hackathons on lablab.ai, participants pull models from the Hub, fine-tune or build on them using AMD Developer Cloud GPUs, and publish their final projects back to the Hub as a Space. Explore what the community has built at Hugging Face Use Cases and Applications.

Hugging Face Hub Tutorials

Getting Started

Key Features

1M+ models Text, vision, audio, multimodal, and specialized domain models from top research teams and companies including Meta, Mistral, Google, and Alibaba Cloud.

Private repositories Host proprietary models and datasets with access controls. Upgrade to a PRO or Enterprise account for private inference endpoints.

Model cards Structured documentation for model limitations, intended use, training details, and evaluation results — standardized across all public checkpoints.

Version control Git-based versioning with LFS support for large files. Every model and dataset on the Hub has a full commit history.

Fine-tuned adapters Share and reuse LoRA and PEFT adapters without uploading full model weights. Adapters reference their base model and load in seconds.

Libraries

  • Transformers Unified API for pretrained models across text, vision, and audio
  • huggingface_hub Python SDK for Hub authentication, uploads, and downloads
  • Datasets Efficient access to Hub datasets with streaming and arrow-based caching
  • PEFT Parameter-efficient fine-tuning (LoRA, QLoRA, prefix tuning)
  • Optimum-AMD Optimized inference and training for AMD hardware via ROCm
Edit on GitHub

huggingface HuggingFace Hub AI technology Hackathon projects

Discover innovative solutions crafted with huggingface HuggingFace Hub AI technology, developed by our community members during our engaging hackathons.

AuraOS

AuraOS

Aura is a sovereign, neuro-symbolic AI operating system designed to make advanced artificial intelligence more transparent, efficient, collaborative, and accountable. Instead of sending every instruction directly to a language model, Aura can locally interpret natural-language intent, organize it into structured semantic and routing representations, identify relevant code or knowledge, and prepare a bounded task before model use is required. Aura combines deterministic routing, a 4,096-primitive lexical system, polysynthetic six-slot intent structures, visual code topology, guarded human-agent workflows, and a Learning Arena where verified experiences can be tested before becoming reusable knowledge. Its Observatory allows users to see how Aura understood an instruction, while the Human Agent Arena supports node-specific dialogue, approval gates, verification, and preserved attempt artifacts—including failed work that may help future refactoring. Aura is built around human authority. Models can propose, explain, investigate, and assist, but they cannot silently expand scope, bypass evidence requirements, or merge changes without approval. The result is an AI architecture focused not only on capability, but on sovereignty, inspectability, learning, and responsible action.

SOAP Copilot: AI Clinical Scribe on AMD

SOAP Copilot: AI Clinical Scribe on AMD

Physicians spend roughly two hours a day on clinical documentation, time pulled directly away from patient care, and a well-documented driver of burnout that costs the US healthcare system an estimated $8.3B a year in lost productivity. SOAP Copilot is a multi-agent AI system that turns a raw doctor-patient conversation into complete, structured clinical documentation in under 30 seconds. Three specialized agents, all built on Llama 3.3 70B Instruct, handle the pipeline in sequence: a SOAP Generator produces a structured Subjective/Objective/Assessment/Plan note, an ICD Coder extracts diagnosis codes with confidence scores, and a Summary Writer rewrites the note as a plain-language patient summary. The entire stack runs on open-source models, so no PHI ever needs to leave the user's own infrastructure. The project was originally prototyped and LoRA fine-tuned on a single AMD MI300X (192GB HBM3) via AMD Developer Cloud, running vLLM on ROCm 7.2. That memory headroom is what makes serving a 70B-parameter model in BF16 precision practical on one chip, something no consumer GPU and few cloud GPUs can do. For this submission, production inference is served through Fireworks AI, which itself runs on AMD Instinct hardware, so the project demonstrates the AMD AI stack end to end: self-managed fine-tuning through to hosted inference. Roadmap items include a distilled specialist model for lower-cost deployment, retrieval-augmented generation over clinical guidelines using the Qdrant vector database already wired into the architecture, and an on-premise, HIPAA-compliant deployment path for health systems that cannot send patient data to any external API.

Thymus

Thymus

Thymus is a lightweight hybrid token-efficient router designed to maximize accuracy while minimizing token costs in multi‑task LLM pipelines. It dynamically routes user queries across local and remote models on LLM providers. The general architecture of the router is as follows. 1. Classify: a fast regex/keyword pass handles the obvious cases for free; anything ambiguous falls back to the local model for classification. 2. Infer locally: the local model (Qwen2.5-3B-Instruct, GGUF, CPU) answers the query and self-reports a routing preference and confidence. 3. Verify: a task-aware verifier sanity-checks the answer (JSON structure, code execution, entity coverage, self-consistency, depending on task type). 4. Decide: LOCAL (free, done), SHEPHERD (local answer used as context, one cheap cloud pass to double check), or RESET (send straight to Fireworks).

Verity — Fact-Checking for the Rest of Us

Verity — Fact-Checking for the Rest of Us

Most people believe first and verify never , there simply isn't time to check every headline, post, or forwarded message before deciding whether to trust it. That gap is where misinformation lives and spreads. Verity closes it. You give Verity a claim — a headline, a statistic, a line from an article — and it separates what's actually checkable from what's just opinion, then searches for independent evidence across 40+ public authoritative sources. Instead of trusting a single model's best guess, Verity requires independent checks to agree before returning a verdict: true, false, or somewhere in between, with a source trail you can click through and verify yourself. Every verification runs live on AMD Instinct hardware via the Fireworks AI API, using Gemma and other open-source models routed through a 5-pass consensus architecture — nothing pre-computed, nothing faked. The models are swappable and inspectable by design, because a trust tool that asks you to trust it blindly defeats its own purpose. Verity is built for the moment right before someone shares something they haven't checked , turning fact-checking from a newsroom luxury into something anyone can do in seconds.

Contxt

Contxt

Every AI grew a memory in 2026 — ChatGPT, Claude, Gemini, Copilot, Grok — but they're five walled gardens that never talk to each other, so you re-introduce yourself every time you switch. And every memory product (Mem0, Supermemory, Letta) is cloud-first and built for developers, not for you. Contxt is the consumer piece nobody built: one portable, privacy-first context layer that works across every AI, where your most sensitive data never leaves your device. A browser extension connects your Gmail, Calendar and Notion, then classifies every item on-device through a Crown-Jewels Gateway — Gemma 3 270M running in WebGPU plus a deterministic-rules safety floor. PRIVATE "crown jewels" stay local and end-to-end encrypted (the cloud is a blind relay of ciphertext, with keys transferred device-to-device by QR). SHARED context is distilled into reusable cards that any AI can read over MCP, and injected live into the composer on Claude, ChatGPT and Gemini. A companion web dashboard mirrors the extension in real time. One context layer. Every AI. Your data stays yours.

AegisLayer: Enterprise Privacy Middleware

AegisLayer: Enterprise Privacy Middleware

AegisLayer is a zero-trust enterprise privacy middleware designed to solve the critical data security bottleneck preventing enterprises from adopting Large Language Models. When employees send prompts to cloud AI providers, sensitive PII and proprietary secrets are often leaked. AegisLayer intercepts these prompts at the network boundary and sanitizes them in real-time before they ever leave the corporate network. Our solution employs a highly optimized dual-engine architecture: 1. CPU Regex Engine: A deterministic pass that instantaneously identifies and redacts structured data like IPv4 addresses, credit card numbers, phone numbers, and API keys. 2. AMD ROCm-Accelerated NER Engine: An advanced Named Entity Recognition pipeline powered by PyTorch and HuggingFace Transformers. Optimized specifically for AMD Instinct GPUs via ROCm, this engine identifies unstructured entities (Persons, Organizations, Locations) with sub-100ms latency. As entities are detected, AegisLayer maps them to opaque, entropy-free tokens (e.g., [PERSON_1]) and stores the mapping in an ephemeral, in-memory vault. The sanitized prompt is then safely forwarded to the external LLM provider. Once the AI generates a response, AegisLayer automatically de-tokenizes the text, restoring the original values seamlessly. The vault is immediately wiped after each round-trip, guaranteeing zero persistent storage of sensitive data. AegisLayer ensures maximum data privacy, regulatory compliance, and architectural resilience without introducing any friction to the end-user experience.