From 9a4f14e704a3e21d9cbd5bf7f4a4e85287ef65f6 Mon Sep 17 00:00:00 2001 From: Georg Grab Date: Thu, 2 Apr 2026 17:09:32 +0200 Subject: [PATCH 1/2] docs: update for HF ungating change --- how-to-access-gated-models.mdx | 68 +++++++++++++--------------------- quickstart.mdx | 2 +- 2 files changed, 26 insertions(+), 44 deletions(-) diff --git a/how-to-access-gated-models.mdx b/how-to-access-gated-models.mdx index b27c474..3665387 100644 --- a/how-to-access-gated-models.mdx +++ b/how-to-access-gated-models.mdx @@ -1,60 +1,42 @@ --- -title: "How to access TabPFN-2.6" +title: "How to access TabPFN-2.5 and TabPFN-2.6" --- -TabPFN-2.6 is released under a [non-commercial license](models#tabpfn-model-license), which you need to accept before the model files can be downloaded. +TabPFN-2.5 and TabPFN-2.6 are released under a [non-commercial license](models#tabpfn-model-license), which you need to accept before the model files can be downloaded. - +## Automatic Setup (recommended) -## Step 1: Accept the license agreement +On first use, TabPFN will automatically open a browser window where you can log in via [PriorLabs](https://ux.priorlabs.ai) and accept the license terms. Your authentication token is cached locally, so you only need to do this once. -Visit Hugging Face, where the model is hosted, and agree to the license terms: https://huggingface.co/Prior-Labs/tabpfn_2_6/ - -You will need to log in to your Hugging Face account, or create one if you don't already have it. Once you have agreed to the license, you should be granted access immediately. - -## Step 2: Create an access token - -- If you created a new account: ensure you have verified your Hugging Face email address -- Create a new token using this link: https://huggingface.co/settings/tokens/new?tokenType=fineGrained - - Give the token a name - - Check the box "Read access to contents of all public gated repos you can access" ( [screenshot](/guide-screenshots/hf-token-creation.png)) -- Press "Create token", and copy the displayed token (you will only be able to see it once).\ - Your token should look like `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` - -## Step 3: Log in with your new token +```python +from tabpfn import TabPFNClassifier -### **If using Google Colab** +# This will open a browser for login and license acceptance on first use +model = TabPFNClassifier() +model.fit(X_train, y_train) +``` -- Select "Secrets" from the left-hand menu ([screenshot](/guide-screenshots/colab-secrets-pane.png)) -- Create a new secret. Set name to "HF_TOKEN" and value to the token itself -- Ensure that "Notebook access" is enabled +## Headless / CI Environments -### **If using a notebook other than Google Colab** +For environments where a browser is not available (e.g. remote servers, CI pipelines, Docker containers): -Ensure TabPFN is installed. Create a new Python code cell containing: +1. Visit [https://ux.priorlabs.ai](https://ux.priorlabs.ai) and log in +2. Go to the **License** tab and accept the license +3. Copy your authentication token from your account +4. Set the `TABPFN_TOKEN` environment variable: -```python -import huggingface_hub -huggingface_hub.login() +```bash +export TABPFN_TOKEN="your-token-here" ``` -Execute this cell and follow the instructions. +You can also set `TABPFN_NO_BROWSER` to disable the automatic browser login if needed (e.g. in environments where opening a browser is undesirable). -### **If using the terminal** +## Google Colab -- Activate the Python environment that you use for TabPFN -- Run `hf auth login` -- Enter the token when prompted +In Google Colab, the browser-based login flow works automatically. Alternatively, you can set the `TABPFN_TOKEN` secret: -### In scripts or CI pipelines - -Configure your environment to specify the `HF_TOKEN` environment variable. +- Select "Secrets" from the left-hand menu +- Create a new secret with name `TABPFN_TOKEN` and value set to your token +- Ensure that "Notebook access" is enabled -For more information, refer to the Hugging Face [documentation](https://huggingface.co/docs/huggingface_hub/en/quick-start#authentication). \ No newline at end of file +If access via the browser-based flow is not an option for you, please contact us at [`sales@priorlabs.ai`](mailto:sales@priorlabs.ai). diff --git a/quickstart.mdx b/quickstart.mdx index 21f446b..e98f59d 100644 --- a/quickstart.mdx +++ b/quickstart.mdx @@ -47,7 +47,7 @@ preds = model.predict(X_test) ``` -The first time you use the model, you will be asked to accept the model license and configure the access token, see the [detailed instructions](/how-to-access-gated-models). +The first time you use the model, a browser window will open for you to log in and accept the model license. For headless environments, see the [detailed instructions](/how-to-access-gated-models). From e4f529dc8258a548c1622ccbfbe2f993263b8869 Mon Sep 17 00:00:00 2001 From: Georg Grab Date: Thu, 2 Apr 2026 17:13:06 +0200 Subject: [PATCH 2/2] fix: address PR review comments on gated models page Use absolute path for license link, remove backticks from email link for style consistency. Co-Authored-By: Claude Opus 4.6 (1M context) --- how-to-access-gated-models.mdx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/how-to-access-gated-models.mdx b/how-to-access-gated-models.mdx index 3665387..790fdcd 100644 --- a/how-to-access-gated-models.mdx +++ b/how-to-access-gated-models.mdx @@ -2,7 +2,7 @@ title: "How to access TabPFN-2.5 and TabPFN-2.6" --- -TabPFN-2.5 and TabPFN-2.6 are released under a [non-commercial license](models#tabpfn-model-license), which you need to accept before the model files can be downloaded. +TabPFN-2.5 and TabPFN-2.6 are released under a [non-commercial license](/models#tabpfn-model-license), which you need to accept before the model files can be downloaded. ## Automatic Setup (recommended) @@ -39,4 +39,4 @@ In Google Colab, the browser-based login flow works automatically. Alternatively - Create a new secret with name `TABPFN_TOKEN` and value set to your token - Ensure that "Notebook access" is enabled -If access via the browser-based flow is not an option for you, please contact us at [`sales@priorlabs.ai`](mailto:sales@priorlabs.ai). +If access via the browser-based flow is not an option for you, please contact us at [sales@priorlabs.ai](mailto:sales@priorlabs.ai).