Hugging Face
Hugging Face is an artificial intelligence technology company that provides a hosting and sharing platform for machine learning models, datasets, and applications (Hugging Face Hub), and maintains open-source machine learning libraries including the widely used Transformers library. The company promotes open-source AI ecosystem development, offering tools for model training, deployment, and collaboration to researchers, developers, and enterprises. It is one of the most influential open-source community platforms in natural language processing and general machine learning.
Core Facts
Hugging Face is the world's largest open-source model platform.
Hugging Face transitioned to an AI open-source community platform in 2018.
Hugging Face's Transformers library supports three major frameworks: PyTorch, TensorFlow, and JAX.
Timeline (last 90 days)
Darwin 模型系列在 Hugging Face 上有公开记录，参考 VIDRAFT 此前发布的曾登顶五个 Hugging Face 基准的 Darwin-180B-RSI，可到其组织主页查看已发布的 checkpoint。
Mistral AI released the Ministral 3 series model weights under the Apache 2.0 license, including two Instruct versions, 3B and 8B, both available on Hugging Face.
HuggingNews is not an official news channel of Hugging Face, but a third-party information aggregation site focused on AI model releases, updates, and community developments, whose content is often cited from Reddit, X, and vendor blogs.
In a Reddit post, the hackathon team explicitly stated that the key reason for choosing the LeRobot solution was that LeRobot is completely open source and supported by Hugging Face, allowing teleoperation, data collection, and imitation learning workflows to be set up quickly.
torch-pointcloud's pre-trained weights are hosted on Hugging Face, and the supported benchmark datasets include ModelNet40, ScanObjectNN, S3DIS, ScanNet, SemanticKITTI, nuScenes, KITTI, and SUN RGB-D.
In the Hugging Face repository JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF, the Q4_K_M quantized file is 8.07 GB in size.
When using the ollama run hf.co/JetBrains/... command, Ollama recognizes the hf.co/ prefix and resolves and downloads the corresponding GGUF directly from Hugging Face rather than pulling from the Ollama registry.
For people following the AI industry, regularly checking the Hugging Face trending list is more effective than waiting for explosive news, because it reflects download counts rather than PR releases.
Closed-source commercial model vendors usually only provide APIs and do not disclose weights, so they cannot directly appear on the Hugging Face trending list.
The strategies of Chinese AI teams include uploading model weights to Hugging Face, publishing technical reports on arXiv, and synchronously open-sourcing GitHub repositories, with no grand launch events or star CEO speaking tours.
40 more timeline events
All Facts (20)
Hugging Face is the world's largest open-source model platform.
90%VerifiedHugging Face transitioned to an AI open-source community platform in 2018.
90%VerifiedHugging Face's Transformers library supports three major frameworks: PyTorch, TensorFlow, and JAX.
90%VerifiedLeRobot is an open-source robot learning framework led by Hugging Face, providing a unified dataset format, pre-trained models, and training pipelines.
80%VerifiedAs of 2023, Hugging Face had completed a Series D funding round with a valuation of $4.5 billion, with investors including Google, Amazon, and Nvidia.
80%VerifiedTGI (Text Generation Inference) is a production-grade inference service launched by Hugging Face, supporting optimization strategies such as continuous batching and tensor parallelism.
80%VerifiedThe Transformers library is maintained by Hugging Face and provides a unified interface for thousands of pre-trained models
80%VerifiedHugging Face is an AI company founded in 2016, whose open-source library Transformers has become the de facto standard toolkit in the NLP and large model domains.
80%VerifiedIn JSONL format, each line is an independent and complete JSON object, supporting streaming reads, parallel processing, and incremental appending; Hugging Face's Datasets library natively supports loading it.
75%VerifiedHugging Face has released a series of supporting open-source libraries including transformers, datasets, and diffusers.
75%VerifiedTransformers.js is a JavaScript library developed by Hugging Face that, based on WebAssembly and WebGPU technology, ports Transformer model inference to browser environments.
70%VerifiedGradio is an open-source Python library maintained by Hugging Face, designed specifically for rapidly building interactive demos for machine learning models.
70%VerifiedQwen-Image-2.1 is available through multiple channels including GitHub, ModelScope, and Hugging Face.
65%VerifiedThe Hugging Face platform has deployed a security.txt file, accessible at https://huggingface.co/security.txt.
65%VerifiedIn the AI industry, a wrapper application typically refers to an application that does not train its own foundation model but instead calls APIs from vendors such as OpenAI, Anthropic, and Google to build a product experience on top of them.
65%VerifiedHugging Face's core open-source library transformers supports the three mainstream frameworks PyTorch, TensorFlow, and JAX
65%VerifiedDeepSeek has made its model weights available on the Hugging Face platform
65%VerifiedCivitai and Hugging Face are platforms where Stable Diffusion-related models can be downloaded for free.
65%VerifiedDiffusers is an open-source Python library developed and maintained by the Hugging Face team, focused on the training, inference, and deployment of diffusion models.
65%VerifiedThe Hugging Face ecosystem includes components such as Model Hub, Datasets, Spaces, and Tokenizers.
65%