model hub
In plain wordsAn online library where trained AI models are published so that others can find, download and reuse them instead of training their own from scratch.
A platform or catalog for publishing, discovering and downloading trained models, typically with versioning, documentation such as model cardsmodel card, n.: A short document published with an AI model explaining what it is for, how it was trained and tested, how well it works for different groups of people, and its known limits., license information, search by task and programmatic download from code. Early collections were called model zoos: curated sets of reference models for one framework, such as Caffe's Model Zoo and later zoos for TensorFlow, PyTorchPyTorch, prop. n.: The most widely used free software toolkit for building and training neural networks, especially in research; programmers write ordinary Python and it handles the calculus and the graphics chips., ONNXONNX, prop. n.: A shared file format for trained AI models, so that a model built with one toolkit can be moved to and run by another, much as a PDF opens on any computer. and MXNet. Open hubs followed in which anyone may upload, above all the Hugging FaceHugging Face, prop. n.: A company whose website is the main public library for sharing AI models and datasets — often called the GitHub of machine learning — and which makes widely used free AI software. Hub, along with TensorFlow Hub (later folded into Kaggle), PyTorch Hub, cloud providers' model catalogs ('model gardens') and the Chinese platform ModelScope. Hubs underpin transfer learningtransfer learning, n.: Reusing what a model learned on one task to get a head start on another — like a pianist picking up the organ faster than a beginner. It lets models learn new jobs from little data. and the open-weightopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private. ecosystem, and they raise supply-chain concerns — malicious files, impersonated organizations, mislabeled licenses and undocumented training data — which they counter with security scanning, gated access, verified organizations and safer formats such as safetensorssafetensors, n.: A file format for storing an AI model's learned numbers that cannot hide harmful programs and loads very quickly, replacing older formats that could run code when opened..
Before fine-tuning anything, search the model hub for a checkpoint already trained on legal text.
Origin
'Model zoo' was popularized by the Caffe framework's shared collection of trained networks (2014). Google launched TensorFlow Hub in 2018 and PyTorch Hub appeared in 2019, the period in which Hugging Face's hub began growing into the dominant open platform.
Usage note
A hub distributes models publicly; a model registrymodel registry, n.: A central catalog where a team stores every version of its trained models, with records of how each was made and which version is approved for use. manages an organization's own versions and releases, though private enterprise hubs combine the two.
Related
- Hugging FaceHugging Face, prop. n.: A company whose website is the main public library for sharing AI models and datasets — often called the GitHub of machine learning — and which makes widely used free AI software.
- model registrymodel registry, n.: A central catalog where a team stores every version of its trained models, with records of how each was made and which version is approved for use.
- model cardmodel card, n.: A short document published with an AI model explaining what it is for, how it was trained and tested, how well it works for different groups of people, and its known limits.
- transfer learningtransfer learning, n.: Reusing what a model learned on one task to get a head start on another — like a pianist picking up the organ faster than a beginner. It lets models learn new jobs from little data.
- open-weight modelopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private.
- safetensorssafetensors, n.: A file format for storing an AI model's learned numbers that cannot hide harmful programs and loads very quickly, replacing older formats that could run code when opened.
- model registrymodel registry, n.: A central catalog where a team stores every version of its trained models, with records of how each was made and which version is approved for use.
Mentioned in
GradioGradio, prop. n.: A free Python library for quickly wrapping an AI model in a simple web page — text boxes, image uploads, chat windows — so that others can try it in a browser.