Keep track of latest changes on the Hugging Face Hub
Organization admins can now invite external collaborators directly from a private repository's settings page. Invitees get read or write access to that repository's Resource Group only.
Resource group admins can also invite external collaborators once an org admin enables "External invitations" in Resource Groups. This setting is off by default and available on Team and Enterprise plans. See the docs for details.
OpenUSD files (.usd, .usda, .usdc and .usdz) can now be viewed on the Hub. Click one in a dataset, model or bucket to open it in an interactive 3D viewer, which loads its layers, payloads and textures. For example, an LG laptop, a Vertiv server rack or a robot arm from BEHAVIOR-1K.
Datasets with SimReady packages also show a 3D Assets preview on their page, with the number of packages from across the dataset, as on LGElectronics/simready-assets and Vertiv/simready-assets.
The viewer is open source and available in huggingface.js as @huggingface/usd-web-viewer.
Sign in to Hugging Face with your fingerprint, face, screen lock or security key instead of your password. Passkeys also work as a second factor, alongside your authenticator app. Add one in your Authentication settings.
The Models and Datasets pages can now be filtered by author, for example huggingface.co/models?author=mistralai. A card above the filters shows whose repos you're browsing, and it stays while you filter by task, size, library, license, etc.
It's accessible from any user or organization profile: the new filter button right before the Sort options, opens the filtered page from there.
The tensor viewer now loads much faster on large sharded models. When you open a model's model.safetensors.index.json, the Hub sends the headers of all shards in a single request, so your browser no longer fetches each shard one by one. For example, DeepSeek-V3.2 now loads in 0.45s instead of 9.2s, GLM-5.3 in 0.57s instead of 4.0s, and Qwen2.5-72B-Instruct in 0.27s instead of 1.2s.
