Update dependency huggingface-hub to v0.36.0 - #91

v0.36.0: [v0.36.0] Last Stop Before 1.0

Compare Source

This is the final minor release before v1.0.0. This release focuses on performance optimizations to HfFileSystem and adds a new get_organization_overview API endpoint.

We'll continue to release security patches as needed, but v0.37 will not happen. The next release will be 1.0.0. We’re also deeply grateful to the entire Hugging Face community for their feedback, bug reports, and suggestions that have shaped this library.

Full Changelog: huggingface/huggingface_hub@v0.35.0...v0.36.0

📁 HfFileSystem

Major optimizations have been implemented in HfFileSystem:

• Cache is kept when pickling a fs instance. This is particularily useful when streaming datasets in a distributed training environment. Each worker won't have to rebuild their cache anymore

• [HfFileSystem] Keep cache on pickle  by @​lhoestq in #​3443

Listing files with .glob() has been greatly optimized:

from huggingface_hub import HfFileSystem

HfFileSystem().glob("datasets/HuggingFaceFW/fineweb-edu/data/*/*")

### Before: ~100 /tree calls (one per subdirectory)
### Now: 1 /tree call

• [HfFileSystem] Optimize maxdepth: do less /tree calls in glob()  by @​lhoestq in #​3389

Minor updates:

• add block_size in init  by @​lhoestq in #​3425
• hffs minor fix  by @​lhoestq in #​3449
• HTTP backoff: Retry on ChunkedEncodingError  by @​lhoestq in #​3437

🌍 HfApi

It is now possible to get high-level information about an organization, the same way it is already possible to do with users:

>>> from huggingface_hub import get_organization_overview
>>> get_organization_overview("huggingface")
Organization(
    avatar_url='https://cdn-avatars.huggingface.co/v1/production/uploads/1583856921041-5dd96eb166059660ed1ee413.png',
    name='huggingface',
    fullname='Hugging Face',
    details='The AI community building the future.',
    is_verified=True,
    is_following=True,
    num_users=198,
    num_models=164, num_spaces=96,
    num_datasets=1043,
    num_followers=64814
)

• Add client support for the organization overview endpoint  by @​BastienGimbert in #​3436

🛠️ Small fixes and maintenance

🐛 Bug and typo fixes

• Add quotes for better shell compatibility  by @​aopstudio in #​3369
• update the sentence_similarity docstring  by @​tolgaakar in #​3374
• Do not retry on 429 (only on 5xx)  by @​Wauplin in #​3377
• Use git xet transfer to check if xet is enabled  by @​hanouticelina in #​3381
• Replace pkgx install instruction with uv  by @​gary149 in #​3420
• The error message as previously displayed...  by @​goldnode in #​3405
• Use all tools unless explicit allowed_tools  by @​Mithil467 in #​3397
• [type validation] skip unresolved forward ref  by @​zucchini-nlp in #​3376
• document job stage possible values  by @​hanouticelina in #​3431
• update token parameter docstring  by @​hanouticelina in #​3447

🏗️ internal

• bump to 0.36.0.dev0 by @​Wauplin (direct commit on main)
• [Workflow] security fix  by @​glegendre01 in #​3383
• migrate tip blocks  by @​hanouticelina in #​3392
• [Internal] Fix ty quality   by @​hanouticelina in #​3441
• backward compatible cli tracking (v0.x)  by @​Wauplin in #​3460

Community contributions

The following contributors have made changes to the library over the last release. Thank you!

• @​aopstudio
      * Add quotes for better shell compatibility (#​3369)
• @​tolgaakar
      * update the sentence_similarity docstring (#​3374) (#​3375)
• @​Mithil467
      * Use all tools unless explicit allowed_tools (#​3397)
• @​goldnode
      * The error message as previously displayed... (#​3405)
• @​BastienGimbert
      * Add client support for the organization overview endpoint (#​3436)

v0.35.3: [v0.35.3] Fix image-to-image target size parameter mapping & tiny agents allow tools list bug

Compare Source

This release includes two bug fixes:

• [Inference] Fix target size mapping for fal-ai's image-to-image in #​3399 by @​hanouticelina flagged by @​iam-tsr
• [Tiny-Agents] Use all tools unless allowed_tools is set explicitly in #​3397 by @​Mithil467

Full Changelog: huggingface/huggingface_hub@v0.35.2...v0.35.3

v0.35.2: [v0.35.2] Welcoming Z.ai as Inference Providers!

Compare Source

Full Changelog: huggingface/huggingface_hub@v0.35.1...v0.35.2

New inference provider! 🔥

Z.ai is now officially an Inference Provider on the Hub. See full documentation here: https://huggingface.co/docs/inference-providers/providers/zai-org.

from huggingface_hub import InferenceClient

client = InferenceClient(provider="zai-org")
completion = client.chat.completions.create(
    model="zai-org/GLM-4.5",
    messages=[{"role": "user", "content": "What is the capital of France?"}],
)

print("\nThinking:")
print(completion.choices[0].message.reasoning_content)
print("\nOutput:")
print(completion.choices[0].message.content)

Thinking:
Okay, the user is asking about the capital of France. That's a pretty straightforward geography question. 

Hmm, I wonder if this is just a casual inquiry or if they need it for something specific like homework or travel planning. The question is very basic though, so probably just general knowledge. 

Paris is definitely the correct answer here. It's been the capital for centuries, since the Capetian dynasty made it the seat of power. Should I mention any historical context? Nah, the user didn't ask for details - just the capital. 

I recall Paris is also France's largest city and major cultural hub. But again, extra info might be overkill unless they follow up. Better keep it simple and accurate. 

The answer should be clear and direct: "Paris". No need to overcomplicate a simple fact. If they want more, they'll ask.

Output:
The capital of France is **Paris**.  

Paris has been the political and cultural center of France for centuries, serving as the seat of government, the residence of the President (Élysée Palace), and home to iconic landmarks like the Eiffel Tower, the Louvre Museum, and Notre-Dame Cathedral. It is also France's largest city and a global hub for art, fashion, gastronomy, and history.

• feat: support zai as inference provider by @​tomsun28 in #​3395

Misc:

• [HfFileSystem] Optimize maxdepth: do less /tree calls in glob() by @​lhoestq in #​3389

v0.35.1: [v0.35.1] Do not retry on 429 and skip forward ref in strict dataclass

Compare Source

• Do not retry on 429 (only on 5xx) #​3377
• Skip unresolved forward ref in strict dataclasses #​3376

Full Changelog: huggingface/huggingface_hub@v0.35.0...v0.35.1

v0.35.0: [v0.35.0] Announcing Scheduled Jobs: run cron jobs on GPU on the Hugging Face Hub!

Compare Source

Scheduled Jobs

In v0.34.0 release, we announced Jobs, a new way to run compute on the Hugging Face Hub. In this new release, we are announcing Scheduled Jobs to run Jobs on a regular basic. Think "cron jobs running on GPU".

This comes with a fully-fledge CLI:

hf jobs scheduled run @​hourly ubuntu echo hello world
hf jobs scheduled run "0 * * * *" ubuntu echo hello world
hf jobs scheduled ps -a
hf jobs scheduled inspect <id>
hf jobs scheduled delete <id>
hf jobs scheduled suspend <id>
hf jobs scheduled resume <id>
hf jobs scheduled uv run @&#8203;weekly train.py

• [Jobs] Add scheduled jobs api by @​lhoestq in #​3306

It is now possible to run a command with uv run:

hf jobs uv run --with lighteval -s HF_TOKEN lighteval endpoint inference-providers "model_name=openai/gpt-oss-20b,provider=groq" "lighteval|gsm8k|0|0"

• [Jobs] Support commands in hf jobs uv run by @​lhoestq in #​3303

Some other improvements have been added to the existing Jobs API for a better UX.

• [Jobs] Use current or stored token in a Job secrets by @​lhoestq in #​3272
• update uv image by @​lhoestq in #​3270

And finally, Jobs documentation has been updated with new examples (and some fixes):

• Fix bash history expansion in hf jobs example by @​nyuuzyou in #​3277
• Add timeout info to Jobs guide docs by @​davanstrien in #​3281
• Update jobs.md by @​tre3x in #​3297
• docs: Add link to uv-scripts organization in Jobs guide by @​davanstrien in #​3326
• docs: Add Docker images section for UV scripts in Jobs guide by @​davanstrien in #​3327
• docs: add link to TRL jobs training documentation by @​davanstrien in #​3330

CLI updates

In addition to the Scheduled Jobs, some improvements have been added to the hf CLI.

• [CLI] print help if no command provided by @​Wauplin in #​3262
• update hf auth whoami output by @​hanouticelina in #​3274
• Add 'user:' prefix to whoami command output for consistency by @​gary149 in #​3267
• Whoami: custom message only on unauthorized by @​Wauplin in #​3288

Inference Providers

Welcome Scaleway and PublicAI!

Two new partners have been integrated to Inference Providers: Scaleway and PublicAI! (as part of releases 0.34.5 and 0.34.6).

• feat: add scaleway inference provider by @​Gnoale in #​3356
• Add PublicAI provider by @​Wauplin in #​3367

Image-to-video

Image to video is now supported in the InferenceClient:

from huggingface_hub import InferenceClient

client = InferenceClient(provider="fal-ai")

video = client.image_to_video(
    "cat.png",
    prompt="The cat starts to dance",
    model="Wan-AI/Wan2.2-I2V-A14B",
)

• [Inference] Support image to video task by @​hanouticelina in #​3289

Miscellaneous

Header content-type is now correctly set when sending an image or audio request (e.g. for image-to-image task). It is inferred either from the filename or the URL provided by the user. If user is directly passing raw bytes, the content-type header has to be set manually.

• [InferenceClient] Add content-type header whenever possible + refacto by @​Wauplin in #​3321

A .reasoning field has been added to the Chat Completion output. This is used by some providers to return reasoning tokens separated from the .content stream of tokens.

• Add reasoning field in chat completion output by @​Wauplin in #​3338

MCP & tiny-agents updates

tiny-agents now handles AGENTS.md instruction file (see https://agents.md/).

• allow use of AGENTS.md as well as PROMPT.md by @​evalstate in #​3317

Tools filtering has already been improved to avoid loading non-relevant tools from an MCP server:

• [MCP] Handle Ollama's deviation from the OpenAI tool streaming spec by @​hanouticelina in #​3140
• [Tiny Agents] Add tools to config by @​NielsRogge in #​3242
• fix allowed tools by @​Wauplin (direct commit on main)

🛠️ Small fixes and maintenance

🐛 Bug and typo fixes

• Fix bad total size after resuming download by @​DKingAlpha in #​3234)
• bug fix: only extend path on window sys by @​vealocia in #​3265
• [Update] HF Jobs Documentation by @​ariG23498 in #​3268
• Improve Git Credential Helper Detection for Linux (GCM & libsecret support) by @​danchev in #​3264
• Make requests decode content by @​rasmusfaber in #​3271
• Add validation warnings for repository limits in upload_large_folder by @​davanstrien in #​3280
• Include HF_HUB_DISABLE_XET in the environment dump by @​hanouticelina in #​3290
• Add type to job owner by @​drbh in #​3291
• Update to use only summary bars for uploads when in notebooks by @​hoytak in #​3243
• Deprecate library/tags/task/... filtering in list_models by @​Wauplin in #​3318
• Added apps as a parameter to HfApi.list_models by @​anirbanbasu in #​3322
• Update error message to improve shell compatibility by @​aopstudio in #​3333
• docs: minor typo fix in /en/guides/manage-cache by @​Manith-Ratnayake in #​3353

🏗️ internal

• Prepare for v0.35 by @​Wauplin in #​3261
• fix-ish CI by @​Wauplin (direct commit on main)
• Fix lfs test in CI by @​Wauplin in #​3275
• [Internal] Use ty type checker by @​hanouticelina in #​3294
• [Internal] fix tycheck quality by @​hanouticelina in #​3320
• Return early in is_jsonable if circular reference by @​Wauplin in #​3348

Community contributions

The following contributors have made changes to the library over the last release. Thank you!

• @​DKingAlpha
  • Fix bad total size after resuming download (#​3234) (#​3248)
• @​vealocia
  • bug fix: only extend path on window sys (#​3265)
• @​danchev
  • Improve Git Credential Helper Detection for Linux (GCM & libsecret support) (#​3264)
• @​rasmusfaber
  • Make requests decode content (#​3271)
• @​nyuuzyou
  • Fix bash history expansion in hf jobs example (#​3277)
• @​tre3x
  • Update jobs.md (#​3297)
• @​hoytak
  • Update to use only summary bars for uploads when in notebooks (#​3243)
• @​anirbanbasu
  • Added apps as a parameter to HfApi.list_models (#​3322)
• @​aopstudio
  • Update error message to improve shell compatibility (#​3333)
• @​Manith-Ratnayake
  • docs: minor typo fix in /en/guides/manage-cache (#​3353)
• @​Gnoale
  • feat: add scaleway inference provider (#​3356)

v0.34.6: [v0.34.6]: Welcoming PublicAI as Inference Providers!

Compare Source

Full Changelog: huggingface/huggingface_hub@v0.34.5...v0.34.6

⚡ New provider: PublicAI

[!Tip]
All supported PublicAI models can be found here.

Public AI Inference Utility is a nonprofit, open-source project building products and organizing advocacy to support the work of public AI model builders like the Swiss AI Initiative, AI Singapore, AI Sweden, and the Barcelona Supercomputing Center. Think of a BBC for AI, a public utility for AI, or public libraries for AI.

from huggingface_hub import InferenceClient

client = InferenceClient(provider="publicai")
completion = client.chat.completions.create(
    model="swiss-ai/Apertus-70B-Instruct-2509",
    messages=[{"role": "user", "content": "What is the capital of Switzerland?"}],
)

print(completion.choices[0].message.content)

• Add PublicAI provider by @​Wauplin in #​3367

v0.34.5: [v0.34.5]: Welcoming Scaleway as Inference Providers!

Compare Source

Full Changelog: huggingface/huggingface_hub@v0.34.4...v0.34.5

⚡ New provider: Scaleway

[!Tip]
All supported Scaleway models can be found here. For more details, check out its documentation page.

Scaleway is a European cloud provider, serving latest LLM models through its Generative APIs alongside a complete cloud ecosystem.

from huggingface_hub import InferenceClient

client = InferenceClient(provider="scaleway")

completion = client.chat.completions.create(
    model="Qwen/Qwen3-235B-A22B-Instruct-2507",
    messages=[
        {
            "role": "user",
            "content": "What is the capital of France?"
        }
    ],
)

print(completion.choices[0].message)

• feat: add scaleway inference provider by @​Gnoale in #​1925

v0.34.4: [v0.34.4] Support Image to Video inference + QoL in jobs API, auth and utilities

Compare Source

Biggest update is the support of Image-To-Video task with inference provider Fal AI

• [Inference] Support image to video task #​3289 by @​hanouticelina

>>> from huggingface_hub import InferenceClient
>>> client = InferenceClient()
>>> video = client.image_to_video("cat.jpg", model="Wan-AI/Wan2.2-I2V-A14B", prompt="turn the cat into a tiger")
>>> with open("tiger.mp4", "wb") as f:
 ...     f.write(video)

And some quality of life improvements:

• Add type to job owner #​3291 by @​drbh
• Include HF_HUB_DISABLE_XET in the environment dump #​3290 by @​hanouticelina
• Whoami: custom message only on unauthorized #​3288 by @​Wauplin
• Add validation warnings for repository limits in upload_large_folder #​3280 by @​davanstrien
• Add timeout info to Jobs guide docs #​3281 by @​davanstrien
• [Jobs] Use current or stored token in a Job secrets #​3272 by @​lhoestq
• Fix bash history expansion in hf jobs example #​3277 by @​nyuuzyou

Full Changelog: huggingface/huggingface_hub@v0.34.3...v0.34.4