AsyncInferenceClient uses synchronous time.sleep

[james-deee]

Describe the bug

We have a highly concurrent environemnt using the async event loop and noticed that everything would get locked up. All of your code was using async, so we went looking in the asyncinferenceclient to see what it did when an inference endpoint wasn't ready......

https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/inference/_generated/_async_client.py#L327

Unfortunately that blocks everything on the event loop. This should instead be:
await asyncio.sleep(1)

Reproduction

No response

Logs

No response

System info

- huggingface_hub version: 0.26.2
- Platform: macOS-14.4.1-arm64-arm-64bit
- Python version: 3.11.7
- Running in iPython ?: No
- Running in notebook ?: No
- Running in Google Colab ?: No
- Running in Google Colab Enterprise ?: No
- Token path ?: /Users/jd/.cache/huggingface/token
- Has saved token ?: True
- Who am I ?: jamie-de
- Configured git credential helpers: osxkeychain
- FastAI: N/A
- Tensorflow: N/A
- Torch: 2.2.0
- Jinja2: 3.1.4
- Graphviz: N/A
- keras: N/A
- Pydot: N/A
- Pillow: 11.0.0
- hf_transfer: N/A
- gradio: N/A
- tensorboard: N/A
- numpy: 1.25.1
- pydantic: 2.9.2
- aiohttp: 3.10.11
- ENDPOINT: https://huggingface.co
- HF_HUB_CACHE: /Users/jd/.cache/huggingface/hub
- HF_ASSETS_CACHE: /Users/jd/.cache/huggingface/assets
- HF_TOKEN_PATH: /Users/jd/.cache/huggingface/token
- HF_STORED_TOKENS_PATH: /Users/jd/.cache/huggingface/stored_tokens
- HF_HUB_OFFLINE: False
- HF_HUB_DISABLE_TELEMETRY: False
- HF_HUB_DISABLE_PROGRESS_BARS: None
- HF_HUB_DISABLE_SYMLINKS_WARNING: False
- HF_HUB_DISABLE_EXPERIMENTAL_WARNING: False
- HF_HUB_DISABLE_IMPLICIT_TOKEN: False
- HF_HUB_ENABLE_HF_TRANSFER: False
- HF_HUB_ETAG_TIMEOUT: 10
- HF_HUB_DOWNLOAD_TIMEOUT: 10

[Wauplin]

Hi @james-deee , thanks for noticing. This is clearly an oversight from our side while implementing the AsyncInferenceClient. I've opened a PR to fix it: #2674

[james-deee]

Thanks! Looks good!