< img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=3131724&fmt=gif" />
Last updated:

    Model Hub

    What is the Model Hub

    The Model Hub is where models are hosted for storage, management, discovery, and sharing. Users can create their own Model Repositories, upload, download, and manage model files, and also explore, retrieve, and use other open models from the hub.

    Core Features

    • Model Upload & Management: Upload model files via the Web interface, Git command line, or SDK, including large files (via Git LFS).
    • Version Control: Manage model files with Git versioning, supporting history browsing and rollbacks.
    • Model Discovery: Search and browse open models by tags, task type, framework, and other dimensions.
    • Access Control: Supports public and private visibility settings; private models are accessible only to authorized users.
    • Model Deployment: One-click deployment of models as dedicated inference instances, or launch fine-tuning jobs.

    Supported Model Formats

    The platform is compatible with mainstream model formats, including:

    • Hugging Face Transformers format (config.json, pytorch_model.bin, .safetensors, etc.)
    • GGUF format (for llama.cpp)
    • Other common model weight files

    Model Card

    Learn what a Model Card is and how to write a well-structured Model Card to describe a model's information, usage, and tags.

    Create Model Repository

    How to create a model repository on the platform, including form field descriptions and post-creation steps.

    Upload Models

    How to upload model files to a model repository via the Web interface, Git, CLI tools, or Python SDK.

    Update Models

    How to edit model files, modify model repository settings, and delete a model repository.

    Download Models

    How to download model files from the Model Hub using Git, CLI tools, or Python SDK.