Change8

HuggingFace Hub

Dev Tools

The official Python client for the Hugging Face Hub.

Latest: v1.29.025 releases10 breaking changes20 common errorsView on GitHub

Release History

View all versions →
v1.29.0Breaking6 fixes5 features
Aug 27, 2026

This release optimizes Xet downloads by caching connection info, introduces the ability to change bucket visibility after creation, and includes several security fixes for path traversal and deserialization vulnerabilities.

v1.28.0Breaking7 fixes7 features
Aug 18, 2026

This release introduces enhanced capabilities for deploying and managing Inference Endpoints, including hardware discovery via CLI and SDK, and support for managed engine images with multi-accelerator parallelism. Several bug fixes and documentation improvements are also included.

v1.27.0Breaking3 fixes5 features
Aug 7, 2026

This release enhances the `hf-cli` with automatic installation and synchronization, introduces flexible engine flag management for Inference Endpoints, and adds Baseten as a new inference provider. It also includes several quality-of-life improvements and bug fixes.

v1.27.0.rc0Breaking4 fixes8 features
Aug 7, 2026

This release enhances the `hf-cli` skill by enabling automatic installation and synchronization, introduces flexible engine flag management for Inference Endpoints, and adds Baseten as a new inference provider. It also includes several quality-of-life improvements and bug fixes.

v1.26.1
Aug 6, 2026
v1.26.0
Jul 30, 2026
v1.26.0.rc2
Jul 30, 2026
v1.26.0.rc1
Jul 30, 2026
v1.26.0.rc0
Jul 30, 2026
v1.25.1
Jul 27, 2026
v1.25.08 fixes2 features
Jul 27, 2026

This release introduces auto-naming for jobs based on their Docker image and command, enhancing job management. It also includes various quality-of-life improvements and bug fixes across different modules.

v1.25.0.rc0
Jul 27, 2026
v1.24.02 fixes5 features
Jul 17, 2026

This release introduces optional naming for Jobs on the Hugging Face Hub, making them easier to find and identify. The documentation has also been significantly updated with a CLI-first approach and new sections for AI agents.

v1.24.0.rc3
Jul 17, 2026
v1.24.0.rc2
Jul 17, 2026
v1.24.0.rc1
Jul 17, 2026
v1.24.0.rc0
Jul 17, 2026
v1.23.03 fixes8 features
Jul 9, 2026

This release introduces the ability to create Hugging Face Spaces from official templates and enhances the CLI with new commands for managing extensions and filtering models. It also improves download progress visibility and offline capabilities for the `hf-cli` skill.

v1.22.0Breaking6 fixes11 features
Jul 3, 2026

This release introduces Sandboxes, isolated cloud machines accessible via API and CLI, and significantly speeds up snapshot downloads by caching repository tree listings. The underlying CLI framework was also migrated from Typer to Click.

v1.20.12 fixes
Jun 25, 2026

Version 1.20.1 addresses minor bugs related to file downloading in both public and private repositories, including fixing a race condition during concurrent downloads.

v1.21.0.rc0Breaking5 fixes4 features
Jun 25, 2026

The Jobs listing functionality was significantly overhauled, introducing server-side filtering and lazy pagination for `list_jobs()`, and renaming the CLI command from `hf jobs ps` to `hf jobs ls`. A critical circular import bug affecting `huggingface_hub.login` was also resolved.

v1.21.0Breaking5 fixes4 features
Jun 25, 2026

This release overhauls the Jobs listing functionality, introducing server-side filtering and lazy pagination for `list_jobs()`, and renames the CLI command from `hf jobs ps` to `hf jobs ls`. A critical bug causing import errors on `login` has also been resolved.

v1.20.0.rc0Breaking2 fixes9 features
Jun 18, 2026

This release introduces browser-based OAuth login as the default authentication method and significantly overhauls large folder uploads using a streamed, multi-commit pipeline powered by Xet. Additionally, Hugging Face Jobs now support waiting for completion, SSH access, and cleaner error reporting.

v1.20.0Breaking2 fixes12 features
Jun 18, 2026

This release introduces browser-based OAuth login as the default authentication method and significantly overhauls folder uploads for speed and reliability using a streamed, multi-commit pipeline. Additionally, Hugging Face Jobs now supports waiting for completion, SSH access, and custom container deployments for Inference Endpoints.

v1.19.0.rc1Breaking3 fixes6 features
Jun 11, 2026

This release introduces secure Keyless CI/CD authentication via OIDC and enhances the CLI by supporting `hf://` URIs for uploads and downloads. It also exposes job container ports through the public jobs proxy and migrates Xet functionality to a new session-based API.

Common Errors

LocalEntryNotFoundError5 reports

This error typically occurs in huggingface-hub. Check the example issues for common solutions.

FileNotFoundError4 reports

FileNotFoundError in huggingface-hub often arises due to temporary network issues during file downloads or corrupted local cache files. Clearing your local cache (using `huggingface-cli cache purge`) and retrying the download is the primary solution. Implement retry mechanisms in your code to automatically handle intermittent network errors, especially when downloading large files.

BadRequestError3 reports

The "BadRequestError" in huggingface_hub often arises when the API receives an improperly formatted request, such as incorrect data types or missing required fields in the input. Inspect your request payload (e.g., 'inputs' parameter) to ensure it adheres to the API's expected schema, typically a dictionary with specific key-value pairs. Double-check your code for instances where None or unexpected objects are being passed as input and correct them according to the API documentation before resubmitting the request.

ReadTimeoutError3 reports

ReadTimeoutError in huggingface-hub usually arises from network instability or server-side delays when downloading or accessing large files/repositories. To fix this, configure the `HF_HUB_DOWNLOAD_TIMEOUT` environment variable with a higher value (in seconds) to allow more time for the download to complete. Alternatively, if using code directly, increase the `timeout` parameter within relevant functions like `hf_hub_download` to an appropriate value for your network conditions.

HfHubHTTPError3 reports

HfHubHTTPError usually indicates a temporary issue on the Hugging Face Hub server (e.g., high traffic, maintenance). Implement retry logic with exponential backoff in your code to handle these transient errors. This allows your program to automatically re-attempt the request after a short delay, increasing the likelihood of success once the server recovers.

HTTPError3 reports

HTTPError in huggingface-hub usually indicates a problem on the Hugging Face Hub's server-side (e.g., server overload, maintenance). Your best course of action is to retry the operation after a short delay (e.g., using exponential backoff) as the server might recover quickly. If the issue persists for an extended period, check the Hugging Face Hub's status page or community forums for updates or report the issue.

Related Dev Tools Packages

Subscribe to Updates

Get notified when new versions are released

RSS Feed