Fact-checked by Grok 2 weeks ago

Hugging Face

Hugging Face is an American-French company and open-source platform that facilitates collaboration in , particularly through its Hugging Face Hub, a repository for sharing models, datasets, and applications across modalities like text, image, audio, and video. Founded in 2016 by Clément Delangue, Julien Chaumond, and Thomas Wolf in , it initially launched as a app aimed at teenagers but pivoted in 2018 to building infrastructure after recognizing the need for accessible tools. The company's mission is to democratize good , one commit at a time, emphasizing open-source development to make advanced accessible to developers, researchers, and organizations worldwide. At the core of Hugging Face's offerings is the Transformers library, a Python package that provides state-of-the-art pretrained models for natural language processing, computer vision, audio, and multimodal tasks, supporting both training and inference with frameworks like PyTorch and TensorFlow. Complementing this are the Datasets library for efficient data loading and processing, and the Hub, which as of October 2025 hosts over 2 million models—more than quadruple the number from early 2024—along with over 500,000 datasets used for tasks ranging from translation to speech recognition. These tools have fostered a vibrant community, with more than 50,000 organizations actively using the platform for AI development and deployment. Hugging Face has grown rapidly, achieving a valuation of $4.5 billion following a $235 million Series D funding round in 2023, backed by investors including , , , and Ventures. By 2025, the company employs around 250 people and generates approximately $130 million in annualized as of 2024, primarily from features like private hubs, compute resources, and inference APIs, while maintaining free access to its core open-source ecosystem. This blend of community-driven innovation and commercial scalability positions Hugging Face as a pivotal force in advancing open , enabling rapid prototyping and deployment of models like and variants.

History

Founding and Early Development

Hugging Face was founded in 2016 in by French entrepreneurs Clément Delangue, who serves as CEO, Julien Chaumond, the CTO, and Thomas Wolf, the Chief Science Officer. The company originated from the founders' shared interest in advancing conversational AI, with Delangue bringing product and marketing expertise, Chaumond contributing engineering and mathematical skills, and Wolf offering scientific and legal insights in AI applications. The initial product was a mobile chatbot application targeted at teenagers, branded as an "AI best friend forever (BFF)" to provide emotional support, entertainment, and interactive companionship beyond traditional productivity tools like . This app leveraged early (NLP) techniques to enable open-domain conversations, aiming to foster engaging interactions through humor and . However, the startup encountered significant early challenges, particularly in sustaining user engagement, as the struggled to maintain long-term interest among its young audience amid the limitations of nascent models at the time. These hurdles were compounded by the company's relocation from to the , where the founders moved to access a larger talent pool and market opportunities in , marking a strategic shift to establish a stronger foothold in the tech ecosystem. To address the technical demands of improving the , the initial team structure emphasized experimentation, with early hires focused on developing and iterating on conversational algorithms using available datasets and models. This small, specialized group empowered rapid prototyping of features, laying the groundwork for deeper exploration into AI-driven dialogue systems despite the engagement obstacles.

Pivot to Machine Learning

In 2018, Hugging Face made a strategic decision to pivot from its initial application to the development and release of open-source (NLP) tools, driven by the transformative potential of the transformer architecture introduced in the 2017 paper "Attention Is All You Need." This shift was further catalyzed by the rapid adoption of models like Google's , released in October 2018, which highlighted the need for accessible implementations in popular frameworks such as . A pivotal moment came when co-founder Thomas Wolf ported to over a single weekend and shared it on , receiving immediate enthusiasm from the community with over 1,000 likes and contributions. This led to the official release of the first version of the Transformers library in late 2018, establishing Hugging Face as a provider of pre-trained models and tools for state-of-the-art tasks. The library quickly gained traction as an open-source resource, reflecting the company's new focus on democratizing through collaborative development. Early community feedback played a crucial role in shaping the library, with users contributing bug fixes, new model integrations, and documentation improvements that drove iterative updates. Hosted on from its inception, the project benefited from the platform's ecosystem, enabling seamless collaboration and version control that accelerated its evolution into a robust toolkit. By , Hugging Face expanded this foundation to include datasets and model sharing capabilities, fostering a collaborative environment for AI practitioners to exchange resources and build upon shared innovations.

Funding, Growth, and Acquisitions

Hugging Face's funding trajectory began to accelerate in late 2019 with a of $15 million led by , enabling expansion of its open-source tools. This was followed by a $40 million Series B in March 2021, led by with participation from and , which supported scaling the Transformers library and community platform. The company's valuation reached $500 million post-Series B, reflecting growing adoption in development. In May 2022, Hugging Face raised $100 million in a Series C round led by , with key investments from and , achieving a $2 billion valuation. A subsequent $235 million Series D in August 2023, led by Ventures and including and , brought total funding to approximately $396 million by 2025. Other prominent backers such as have consistently supported the company's focus on collaborative AI infrastructure. These investments fueled rapid growth, with employee numbers growing to around 160 by 2023 and approximately 250 by 2025, alongside a valuation climbing to $4.5 billion. Strategic acquisitions have complemented this expansion. In December 2021, Hugging Face acquired Gradio, a library for creating customizable user interfaces for models. In June 2024, it acquired Argilla, a platform for collecting and managing human feedback in development. In August 2024, Hugging Face acquired XetHub, a Seattle-based startup specializing in scalable for models, to enhance collaboration on large datasets. The most notable move came in April 2025 with the acquisition of Pollen Robotics, a humanoid firm, for an undisclosed amount, aimed at integrating with software. This deal enabled the release of the SO-101, a 3D-printable starting at $100, designed for accessible experimentation in -driven .

Core Technologies

Transformers Library

The Transformers library is an open-source library developed by Hugging Face that serves as a unified for accessing, loading, and utilizing state-of-the-art transformer-based models across domains such as , , audio, video, and multimodal tasks. It emphasizes ease of use by providing model definitions that are compatible with major frameworks, including as the primary backend, alongside and through dedicated support and converters. Initially released on November 17, 2018, the library has undergone continuous development, reaching version 4.57.1 by October 2025, with regular updates incorporating new architectures and optimizations. A core strength of the Transformers library lies in its pipeline API, which abstracts complex model loading and inference into simple, task-oriented interfaces for applications like text classification, , , and . This enables users to perform high-level operations with minimal code, automatically handling preprocessing, model execution, and postprocessing. The library supports over 300 distinct architectures, encompassing encoder-only models like for bidirectional text representation, decoder-only models such as variants for autoregressive generation, encoder-decoder setups for sequence-to-sequence tasks, and multimodal extensions including CLIP for cross-modal alignment of text and images, as well as Vision Transformers for patch-based visual feature extraction. Internally, it manages transformer-specific components like tokenization via fast Rust-based preprocessors tailored to each architecture and efficient attention mechanisms, ensuring compatibility and performance across models. To illustrate practical usage, the library allows quick instantiation of pre-trained models for inference, as shown in the following example for :
python
from transformers import [pipeline](/page/Pipeline)

classifier = [pipeline](/page/Pipeline)("sentiment-analysis")
result = classifier("I love using Hugging Face!")
print(result)  # Outputs: [{'label': 'POSITIVE', 'score': 0.9998}]
This loads a default pre-trained model, processes input text, and returns predictions with confidence scores, leveraging automatic tokenization and model execution under the hood. Since its inception, the Transformers library has evolved to include robust fine-tuning tools, such as the Trainer class, which streamlines workflows with built-in support for distributed training, gradient accumulation, and evaluation metrics. Optimizations have been integrated for transformer-specific challenges, including accelerated computations via FlashAttention to reduce usage and computation time during both training and inference, as well as handling of tokenizer configurations that adapt to diverse languages and modalities. These enhancements have made the library suitable for large models on standard hardware, focusing on conceptual rather than exhaustive hardware specifics. In terms of performance, the library incorporates features like model parallelism—encompassing , tensor parallelism, and parallelism—to distribute computation across multiple devices, enabling the training and inference of models too large for single GPUs. Benchmarks highlight substantial speedups; for instance, integration with tools like DeepSpeed for optimization can yield 2-10x reductions in and training time for billion-parameter models compared to baseline implementations, depending on scale and hardware configuration. Such capabilities underscore the library's role in democratizing access to high-performance models while maintaining focus on core architectural efficiency.

Supporting Libraries

The Hugging Face ecosystem includes several supporting libraries that facilitate data preparation, tokenization, distributed training, efficient , and specialized model handling, enabling seamless workflows beyond core model . These libraries are designed to integrate tightly with the broader platform, allowing users to load datasets, preprocess inputs, scale training across hardware, and apply advanced techniques like parameter-efficient adaptation, all while leveraging the Hugging Face Hub for sharing resources. The Datasets library provides tools for easily loading, processing, and sharing AI datasets across , , and audio tasks. It supports streaming large datasets directly from the , which is particularly useful for handling multi-terabyte collections without full downloads, as demonstrated in recent optimizations for prefetching and buffering introduced in late 2025. By November 2025, the library enables access to over 544,000 datasets hosted on the , including examples like the FineVision dataset with 24 million image-text pairs for vision-language model training. Key features include built-in , such as random cropping or text perturbations, and support for data formats that combine text, images, and audio for diverse applications. The Tokenizers library offers fast, customizable tokenization algorithms tailored for various languages and model architectures. It implements efficient methods like Byte-Pair Encoding (BPE), which merges frequent character pairs to build subword vocabularies, reducing out-of-vocabulary issues in multilingual settings. This library processes text into tensor inputs optimized for transformer models, with Rust-based backends ensuring high performance even on large corpora. Other prominent libraries include Accelerate, which simplifies distributed training by allowing the same code to run across single GPUs, multiple GPUs, TPUs, or clusters with minimal modifications—typically just four lines of code for setup. PEFT (Parameter-Efficient ) enables methods like Low-Rank Adaptation (), which fine-tunes large models by updating only a small subset of parameters, drastically reducing memory and compute needs while maintaining performance. Diffusers specializes in pretrained models for generating images, videos, and audio, providing pipelines for tasks like text-to-image synthesis with easy customization. These libraries interoperate closely with the Transformers library; for instance, the Datasets library can stream and preprocess data directly into training loops managed by Accelerate, while PEFT adapters apply to models loaded via Transformers for efficient . This integration streamlines end-to-end workflows, from data ingestion to optimized training. By 2025, recent additions have expanded support for advanced techniques, including the GRPO (Group Relative Policy Optimization) trainer in the TRL (Transformer Reinforcement Learning) library, which facilitates (RLHF) through online iterative improvements using self-generated data. Additionally, enhancements in Datasets and Diffusers have bolstered tools for audio and vision tasks, such as multimodal streaming for vision-language datasets and diffusion-based audio generation pipelines.

Safetensors

Safetensors is a developed by Hugging Face that provides a secure and efficient for model weights, serving as a safer alternative to PyTorch's to mitigate vulnerabilities such as during model loading. This addresses critical security risks in shared model repositories, where malicious code embedded in files could compromise user systems upon deserialization. Key features of Safetensors include deserialization, which allows tensors to be loaded directly into memory without intermediate copying, enabling faster inference startup times. It supports tensors from multiple frameworks, including , , , and , through and bindings that facilitate seamless integration. The file format consists of a compact 8-byte header indicating the size of the , followed by a JSON-encoded header containing tensor details such as names, data types (e.g., bfloat16, fp8), shapes, and byte offsets, and then the raw binary tensor data stored in little-endian, row-major order without striding. This structure supports sharded files for large models, avoiding file size limits and enabling in distributed environments. Safetensors was released in September 2022 and quickly integrated into the Hugging Face Transformers library and , becoming the recommended standard for uploading models to prevent risks associated with formats. By 2025, nearly all new models on the Hugging Face , including major releases like , Gemma, and , are stored in the Safetensors format. Performance benchmarks demonstrate Safetensors' efficiency: for the BLOOM model, loading times were reduced from 10 minutes using pickle to 45 seconds on 8 GPUs, representing over 13x in this case. On CPU, loading is extremely fast compared to pickle, while GPU loading matches or exceeds equivalents, with general improvements of 2-5x for typical models like GPT-2. In 2025, Safetensors received enhancements for better support of quantized models, including compatibility with formats like GPTQ and AWQ for reduced precision weights, and improved sharding for multi-GPU deployments. These updates also facilitate integration with enterprise security protocols, such as secure model catalogs that scan for vulnerabilities in distributed AI environments.

Platform and Services

Hugging Face Hub

The Hugging Face Hub is a central collaborative platform for the community, functioning as a Git-based that enables hosting, discovery, and versioning of resources such as models and sets. Launched in 2019, it has grown significantly, hosting over 2 million models, more than 500,000 sets, and over 1 million interactive demos called Spaces as of 2025. This infrastructure democratizes access to pre-trained models and , allowing users to share and build upon open-source contributions without proprietary barriers. Key features of the Hub include model cards, which provide comprehensive metadata for each hosted model, such as usage instructions, supported tasks, languages, ethical considerations, potential biases, and limitations. Similarly, dataset viewers facilitate exploration through Dataset Cards and the Data Studio, enabling interactive previews and analysis of structured data. is powered by , with support for Git LFS to handle large files efficiently, allowing users to track changes via commit histories, diffs, and branches. Collaboration is streamlined through familiar tools like forking repositories, submitting pull requests for contributions, and participating in discussions directly on the platform. The Hub integrates with , enabling seamless synchronization of repositories and broader code-sharing workflows. For search and discovery, users can apply filters by task (e.g., text classification or image generation), supported library (e.g., Transformers), and language, while trending sections highlight popular and recently updated resources to aid navigation across the vast collection. Spaces extend the Hub's utility by offering no-code hosting for interactive applications, primarily built using Gradio or Streamlit SDKs. These allow creators to deploy demos for diverse tasks, such as building chatbots for interaction or tools for image generation and editing, with over 1 million public Spaces available for experimentation and reuse.

Inference and Deployment Tools

Hugging Face provides a suite of tools designed to facilitate the inference and deployment of models in production environments, enabling developers to run models at scale without managing underlying infrastructure. These tools bridge the gap between model development on the Hugging Face Hub and real-world applications, supporting everything from quick prototyping to high-throughput serving. Central to this ecosystem is the emphasis on accessibility, optimization, and integration with major cloud platforms. The Inference API offers a serverless solution for rapid model testing, allowing users to perform inference via simple HTTP endpoints on thousands of models hosted on the Hugging Face Hub without any setup or infrastructure management. It includes a free tier suitable for experimentation, with rate limits that scale for subscribers, and supports tasks such as , , and audio processing through a unified or client. This API is particularly useful for validating model performance in low-stakes scenarios, powering interactive playgrounds where users can query models directly in the . For production-grade deployment, Inference Endpoints enable the hosting of dedicated, scalable instances of models on GPU, CPU, or accelerator hardware, with pay-as-you-go pricing starting at $0.033 per hour for basic CPU cores and $0.50 per hour for entry-level GPUs like T4 as of 2025. Users can configure auto-scaling by setting minimum and maximum replicas to handle variable loads, and select custom hardware options across providers such as AWS, , and , including advanced instances like A100 GPUs or AWS Inferentia2 chips. This service ensures low-latency responses and secure, isolated environments, billed per minute of active compute usage. Complementing these deployment options, the Optimum library extends the Transformers framework to optimize models specifically for efficient inference, incorporating techniques like ONNX Runtime export for cross-platform compatibility and quantization methods that reduce model size and accelerate execution on diverse hardware. For instance, 8-bit or 4-bit quantization can yield up to 4x speedups in latency while maintaining accuracy, making it ideal for resource-constrained settings. Optimum integrates seamlessly with pipelines for tasks like or summarization, allowing developers to and run optimized models via a single call. Hugging Face's tools integrate natively with leading providers to simplify scaling and serverless deployment. On AWS, models can be deployed via endpoints using dedicated SDK extensions that handle and automatically. Similarly, Google Cloud integration supports deployment on Engine (GKE) or Vertex AI for managed inference, enabling low-latency applications through serverless options like Cloud Run. These integrations allow for hybrid setups, where models from the are pulled directly into workflows for seamless orchestration. In 2025, Hugging Face enhanced its capabilities with a focus on edge deployment for mobile, , and applications, bolstered by the April acquisition of Pollen Robotics. This move integrated open-source like the , featuring a mobile base with for navigation, into the LeRobot platform, which provides PyTorch-based tools for on-device model training and in real-world embodied scenarios. These advancements lower barriers for deploying optimized models on edge devices, tying software optimizations from Optimum to physical for applications in autonomous systems and teleoperated .

Enterprise Offerings

Hugging Face provides enterprise-grade solutions through its Enterprise Hub, which enables organizations to privately host and collaborate on models, datasets, and applications with enhanced and tools. Key features include unlimited repositories, role-based controls via Groups, and with (SSO) protocols such as SAML and SCIM for provisioning. for the Enterprise Hub starts at $50 per per month, with options for annual commitments and managed billing to support scalable team deployments. Complementing the Hub, AutoTrain offers a no-code for custom models, supporting supervised tasks like and , as well as tasks such as clustering. Enterprise users can leverage AutoTrain Spaces within the Hub for seamless, GPU-accelerated training without infrastructure management, making it suitable for and deployment of tailored solutions. This service abstracts complex training pipelines, allowing businesses to iterate on models using their proprietary data while maintaining . Hugging Face's professional services include dedicated expert support for model customization and optimization consulting, helping enterprises integrate AI into production workflows. These services facilitate partnerships with major players like and , enabling collaborative development of customized large language models and deployment strategies. For instance, integrations with IBM's watsonx and Salesforce's Einstein platforms allow for secure, scalable AI applications built on open-source foundations. Security is a of these offerings, with the Enterprise Hub achieving SOC 2 Type 2 compliance and GDPR adherence to ensure data protection and auditability. Features encompass audit logs for tracking model usage, scanning on uploads, and private endpoints for Endpoints to isolate sensitive computations. These measures support regulatory requirements and mitigate risks in enterprise AI deployments. In 2025, following the April acquisition of Pollen Robotics, Hugging Face expanded its enterprise services to include hardware integration for and edge applications. This move introduces support for deploying open-source models on robots like Reachy 2, enabling businesses to customize edge deployments with optimized hardware-software stacks for real-world tasks.

Community and Impact

Open-Source Ecosystem

Hugging Face's open-source ecosystem is built around a vast collaborative , comprising over five million registered users as of 2025, who actively contribute to the and refinement of models, datasets, and applications. This scale is evidenced by more than two million public models hosted on the platform, alongside hundreds of thousands of datasets and spaces created by contributors worldwide. The engages through regular events, such as Community Weeks focused on specific technologies like and for and tasks, fostering hands-on collaboration and knowledge sharing among participants. Contributions operate under an open governance model primarily hosted on , where repositories like Transformers encourage pull requests, issue discussions, and code reviews from global developers to iteratively improve libraries and models. To incentivize high-impact work, Hugging Face offers bounties via GitHub issues and grants through programs like the Fellowship, which supports early-career researchers in advancing open projects. Key initiatives underscore this collaborative spirit; for instance, the BigScience workshop from to 2022 united over 1,000 researchers to develop the BLOOM multilingual , emphasizing transparent training processes and . Complementing such efforts, ethical guidelines are integrated into model cards, requiring creators to document intended uses, biases, limitations, and societal impacts to promote responsible development. Collaboration is facilitated by built-in tools like discussion forums for peer feedback and leaderboards that benchmark model performance on standards such as GLUE and SuperGLUE, enabling competitive yet cooperative advancements in . These features allow users to compare results, share insights, and build upon each other's work without proprietary barriers. To address inclusivity, Hugging Face runs diversity-focused programs, including the AI Research Residency and Fellowship initiatives, which prioritize applicants from underrepresented groups in to broaden participation and perspectives in the ecosystem.

Adoption and Broader Influence

Hugging Face's tools and platform have achieved broad industry adoption, powering AI initiatives for over 50,000 organizations worldwide, including major enterprises in technology, finance, and healthcare. In natural language processing, companies deploy Hugging Face models to build intelligent chatbots that handle customer interactions with high accuracy and scalability, while in computer vision, they enable applications like object detection in manufacturing quality control. Generative AI use cases, such as content creation and image synthesis, further demonstrate its versatility, with businesses fine-tuning models like Stable Diffusion for customized creative workflows. A key example of enterprise integration is Hugging Face's partnership with , where models from the are seamlessly incorporated into the watsonx.ai to support scalable deployments in and . For sentiment analysis at scale, organizations fine-tune BERT-based models to process vast customer feedback datasets, improving market insights without requiring extensive in-house expertise. These applications highlight how Hugging Face reduces development time and costs, allowing teams to focus on rather than foundational . The platform's broader influence stems from its role in democratizing AI, providing free access to pre-trained models, datasets, and tutorials that lower barriers for developers and researchers globally. This accessibility has accelerated AI research, with the Transformers library serving as a foundation for numerous state-of-the-art advancements, evidenced by over 20 billion downloads of top models on the . By fostering an open ecosystem, Hugging Face has influenced ethical AI practices through transparent model sharing via model cards, which document biases, limitations, and usage guidelines to promote responsible deployment. However, the platform has faced challenges with security, including the identification of over 100 malicious models in early 2025 that exploited pickle file vulnerabilities for potential code execution; Hugging Face responded swiftly by removing the models and improving scanning tools like Picklescan. In emerging areas, Hugging Face's April 2025 acquisition of Pollen Robotics marks a significant push into AI-enabled robotics, open-sourcing designs for humanoid robots like Reachy 2 to integrate large language models with physical actions. This initiative includes hardware innovations such as 3D-printed arms, enabling customizable, affordable robotics for research and applications in automation and human-robot interaction. Following the acquisition, Hugging Face launched the Reachy Mini, an open-source desktop humanoid robot in July 2025, priced starting at $299 for the lite version, to facilitate broader experimentation with AI-driven robotics. Overall, these efforts address key challenges by making advanced AI and robotics accessible to non-experts, while emphasizing transparency to mitigate ethical risks in deployment.

References

  1. [1]
    Hugging Face – The AI community building the future.
    The AI community building the future. The platform where the machine learning community collaborates on models, datasets, and applications. Explore AI Apps.About · Spaces · Log In · Welcome - Hugging Face
  2. [2]
    Hugging Face | Company Overview & News - Forbes
    Founded by Clement Delangue, Julien Chaumond and Thomas Wolf, Hugging Face has some 10 million users who use it to share code and collaborate on models ...
  3. [3]
    Report: Hugging Face Business Breakdown & Founding Story
    Clément Delangue (CEO) and Julien Chaumond (CTO) founded Hugging Face in 2016 as an “AI best friend forever (BFF)” chatbot via a mobile app for teenagers. The ...
  4. [4]
    huggingface (Hugging Face)
    We are on a mission to democratize good machine learning, one commit at a time. If that sounds like something you should be doing, why don't you join us!
  5. [5]
    Transformers
    ### Summary of Transformers Library
  6. [6]
    Announcing New Dataset Search Features - Hugging Face
    Jul 8, 2024 · The AI and ML community has shared more than 180,000 public datasets on The Hugging Face Dataset Hub. Researchers and engineers are using these ...
  7. [7]
    [PDF] Model Card Metadata Collection from Hugging Face to ... - SciTePress
    As of March 2025, Hugging Face hosts more than 1.4 million models, which were 0.5 million approximately in February 2024. In this dataset paper, we provide ...Missing: statistics | Show results with:statistics
  8. [8]
    AI Startup Hugging Face Valued at $4.5B - Bloomberg.com
    Aug 24, 2023 · Hugging Face, a startup that makes AI software and hosts it for other companies, said it has been valued at $4.5 billion after raising $235 ...<|control11|><|separator|>
  9. [9]
    Hugging Face revenue, valuation & growth rate - Sacra
    ... Hugging Face lets users host unlimited models, datasets, and spaces, public or private, for free. The vast majority of their ~$70M in annualized revenue ...
  10. [10]
    Hugging Face: Open-Sourcing the Future of AI | Sequoia Capital
    Jun 29, 2023 · That's when the mission at Hugging Face changed. Delangue and his co-founders decided to start sharing everything that they had learned about ML ...
  11. [11]
    What is Hugging Face? - IBM
    Hugging Face was founded by French entrepreneurs Clément Delangue, Julien Chaumond, and Thomas Wolf in New York City in 2016.1 The entrepreneurs were originally ...
  12. [12]
    Video and transcript: Fireside chat with Clem Delangue, CEO of ...
    Mar 29, 2023 · Clem is the CEO and co founder of Hugging Face, which is really one of the main pieces of infrastructure that everybody uses in the AI industry.
  13. [13]
  14. [14]
    Clem Delangue, Cofounder/CEO HuggingFace (TRANSCRIPT)
    Co-hosts Sarah Guo and Elad Gil talk to Clem Delangue, co-founder and CEO of Hugging Face. Clem shares how they shifted away from their original product, ...
  15. [15]
    Releases · huggingface/transformers - GitHub
    Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference ...Missing: pivot 2018
  16. [16]
    Hugging Face raises $15 million to build the definitive ... - TechCrunch
    Dec 17, 2019 · Hugging Face has raised a $15 million funding round led by Lux Capital. The company first built a mobile app that let you chat with an artificial BFF.
  17. [17]
    Hugging Face raises $40 million for its natural language processing ...
    Addition is leading the round. The company has been building an open source ...
  18. [18]
    How Much Did Hugging Face Raise? Funding & Key Investors - Clay
    Mar 24, 2025 · Angel Round · : $1.2M · : March 2017 ; Seed Round · : $4M · : May 2018 ; Series A · : $15M · : December 2019 ; Series B · : $40M · : March 2021 ; Series C.
  19. [19]
    Hugging Face nabs $100M to build the GitHub of machine learning
    May 9, 2022 · It's a $100 million Series C round with a big valuation. Following today's funding round, Hugging Face is now worth $2 billion. Lux Capital is ...
  20. [20]
  21. [21]
    Hugging Face 2025 Company Profile: Valuation, Funding & Investors
    Information on valuation, funding, cap tables, investors, and executives for Hugging Face. Use the PitchBook Platform to explore the full profile.
  22. [22]
    Hugging Face acquires Seattle data storage startup XetHub
    Aug 8, 2024 · Hugging Face has acquired XetHub, a data storage and collaboration startup founded by former Apple engineers that helped developers streamline the process.
  23. [23]
    Hugging Face buys a humanoid robotics startup | TechCrunch
    Apr 14, 2025 · AI dev platform Hugging Face has acquired Pollen Robotics, a robotics startup based in France, for an undisclosed amount.Missing: Nomic 2021 2024
  24. [24]
    Hugging Face releases a 3D-printed robotic arm starting at $100
    Apr 28, 2025 · AI startup Hugging Face has released a 3D-printed robotic arm for $100, building on its existing robotics efforts.
  25. [25]
    Libraries - Hugging Face
    Natural language processing library built on top of Keras that works natively with TensorFlow, JAX, or PyTorch.
  26. [26]
    The Transformers Library: standardizing model definitions
    May 15, 2025 · Transformers is the default library for LLMs and VLMs in the Python ecosystem. Transformers now supports 300+ model architectures, with an average of ~3 new ...
  27. [27]
    Auto Classes
    Summary of each segment:
  28. [28]
    Safetensors - Hugging Face
    Safetensors is a new simple format for storing tensors safely (as opposed to pickle) and that is still fast (zero-copy). Safetensors is really fast.Convert weights to safetensors · Torch API · Tensorflow API · PaddlePaddle APIMissing: features | Show results with:features
  29. [29]
    huggingface/safetensors: Simple, safe way to store and ... - GitHub
    This repository implements a new simple format for storing tensors safely (as opposed to pickle) and that is still fast (zero-copy).Issues 23 · Actions · Pull requests 4 · Discussions
  30. [30]
    Secure Deserialization of Pickle-based Machine Learning Models
    Aug 21, 2025 · Hugging Face released the SafeTensors format in September 2022 as an alternative that prioritizes security (Hugging Face, 2024b) . The GGUF ...
  31. [31]
    Common AI Model Formats - Hugging Face
    Feb 27, 2025 · New models released on Hugging Face are all stored in safetensors format, including Llama, Gemma, Phi, Stable-Diffusion, Flux, and many others.
  32. [32]
    Load safetensors - Hugging Face
    This guide will show you how you load .safetensor files, and how to convert Stable Diffusion model weights stored in other formats to .safetensor.Missing: structure | Show results with:structure
  33. [33]
    Speed Comparison - Hugging Face
    Safetensors is really fast. Let's compare it against PyTorch by loading gpt2 weights. To run the GPU benchmark, make sure your machine has GPU or you have ...
  34. [34]
    Safetensors: The Secure, Scalable Format Powering LLM Inference
    Apr 23, 2025 · Quantized & Sharded Support: Easily supports formats like GPTQ or AWQ across multiple GPUs. Almost every major open-source model is published ...
  35. [35]
    Secure LLM Models Catalog 2025 - Christian Taillon
    Oct 23, 2025 · Security scan links are provided for each model. Security scans look for malicious code, unsafe weights, backdoors, and other security concerns.
  36. [36]
    Hugging Face Hub documentation
    The Hub is home to over 500k public datasets in more than 8k languages that can be used for a broad range of tasks across NLP, Computer Vision, and Audio. The ...The Model Hub · Datasets · Datasets Overview · Academia HubMissing: November 2025
  37. [37]
    huggingface_hub v1.0: Five Years of Building the Foundation of ...
    Oct 27, 2025 · TL;DR: After five years of development, huggingface_hub has reached v1.0 - a milestone that marks the library's maturity as the Python ...
  38. [38]
    Model Cards - Hugging Face
    The model card should describe: the model; its intended uses & potential limitations, including biases and ethical considerations as detailed in Mitchell, 2018 ...Create and share Model Cards · Annotated Model Card Template
  39. [39]
  40. [40]
  41. [41]
    Inference Providers - Hugging Face
    Hugging Face's Inference Providers give developers access to hundreds of machine learning models, powered by world-class inference providers.HF Inference · Inference Providers on the Hub · Pricing and Billing · Fal AIMissing: acquisition 2024
  42. [42]
    Inference - Hugging Face
    InferenceClient aims to provide a unified experience to perform inference. The client can be used seamlessly with either the (free) Inference API, self-hosted ...
  43. [43]
    Pricing - Hugging Face
    When you create an Endpoint, you can select the instance type to deploy and scale your model according to an hourly rate. Inference Endpoints is accessible to ...Pricing
  44. [44]
    Optimum - Hugging Face
    🤗 Optimum is an extension of Transformers that provides performance optimization tools to train and run models on targeted hardware with maximum efficiency.Optimum Intel · Optimum-AMD · Quantization · Installation
  45. [45]
    Accelerated Inference with Optimum and Transformers Pipelines
    May 10, 2022 · Optimum can be used for accelerated training, quantization, graph optimization, and now inference as well with support for transformers pipelines.3. End-To-End Tutorial On... · 3.2 Convert A Hugging Face... · 3.6 Evaluate The Performance...
  46. [46]
    Hugging Face on AWS
    Amazon SageMaker SDK provides a seamless integration specifically designed for Hugging Face models, simplifying the deployment process of managed endpoints.
  47. [47]
    Hugging Face on Google Cloud
    Hugging Face enables new experiences for Google Cloud customers. They can easily train and deploy Hugging Face models on Google Kubernetes Engine (GKE) and ...
  48. [48]
    Hugging Face to sell open-source robots thanks to Pollen Robotics ...
    Apr 14, 2025 · We're excited to take it a step further by welcoming Pollen Robotics to Hugging Face, a team that's spent the last 9 years building open-source robots and ...Missing: Nomic 2021
  49. [49]
    LeRobot - Hugging Face
    LeRobot aims to provide models, datasets, and tools for real-world robotics in PyTorch. The goal is to lower the barrier for entry to robotics.Lerobot/smolvla_base · Lerobot/diffusion_pusht · Aloha_mobile_cabinet · PapersMissing: edge deployment mobile IoT 2025
  50. [50]
    Pricing - Hugging Face
    Starting at $0.033/hour. Inference Endpoints (dedicated) offers a secure production solution to easily deploy any ML model on dedicated and autoscaling ...
  51. [51]
    Team & Enterprise plans - Hugging Face
    In this section we will document the following Enterprise Hub features: Single Sign-On (SSO) · Advanced Single Sign-On (SSO) · User Provisioning (SCIM) · Audit ...
  52. [52]
    AutoTrain - Hugging Face
    AutoTrain is the perfect tool for anyone eager to dive into the world of machine learning without getting bogged down by the complexities of model training.Finetune Mixtral 8x7B with... · Extractive Question Answering... · Train Locally
  53. [53]
    Enterprise Hub Cookbook - Hugging Face Open-Source AI Cookbook
    With AutoTrain Spaces, you can train your own machine learning models in a simple interface without any code. Read this recipe to learn how to fine-tune your ...
  54. [54]
    Expert Support - Hugging Face
    Talk to Sales to discuss your project and requirements. Unlock priority support for Enterprise Hub and build better AI in-house with Hugging Face Experts.
  55. [55]
    IBM and Salesforce Expand Partnership to Advance Open, Trusted ...
    May 21, 2024 · IBM and Salesforce are working together to make it easier to navigate the complex generative AI landscape by helping clients select the right ...<|control11|><|separator|>
  56. [56]
    Security - Hugging Face
    Hugging Face is also SOC2 Type 2 certified, meaning we provide security certification to our customers and actively monitor and patch any security weaknesses.User access tokens · Git over SSH · Single Sign-On (SSO) · Malware Scanning
  57. [57]
    Enterprise Hub - Hugging Face
    The most advanced platform to build AI with enterprise-grade security, access controls, dedicated support and more.Enterprise HubBilling
  58. [58]
    Hugging Face: AI hub with 5M users, 1M models, and $70M revenue
    Aug 30, 2025 · With over five million users and one million+ open-source models, Hugging Face has become the go-to hub for AI engineers, researchers, ...Missing: statistics | Show results with:statistics
  59. [59]
    [Open-to-the-community] Community week using JAX/Flax for NLP ...
    Jun 23, 2021 · ... Hugging Face, and the open-source NLP & CV community. The talks will take place before the community week from June 30th to July 2nd. Make ...
  60. [60]
    Announcing the Hugging Face Fellowship Program
    May 17, 2022 · The application deadline is June 13, 2022. Hugging Face is actively working to build a culture that values ​​diversity, equity, and inclusion.How The Fellowship Works... · How To Become A Fellow · Frequently Asked Questions
  61. [61]
    BigScience Research Workshop
    It's an open collaboration boot-strapped by HuggingFace, GENCI and IDRIS, and organised as a research workshop. This research workshop gathers academic, ...Bloom · 5 (ACL 2022) · Blog
  62. [62]
    SuperGLUE - a Hugging Face Space by evaluate-metric
    SuperGLUE (https://super.gluebenchmark.com/) is a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, ...
  63. [63]
    Announcing the AI Research Residency Program - Hugging Face
    Mar 22, 2022 · We are actively working to build a culture that values diversity, equity, and inclusivity. We are intentionally building a workplace where ...
  64. [64]
    Fine-tune small models with LLM insights - a CFM case study
    Dec 3, 2024 · This article presents a deep dive into Capital Fund Management's (CFM) use of open-source large language models (LLMs) and the Hugging Face (HF) ecosystem.Missing: adoption studies
  65. [65]
    Hugging Face and IBM partner on watsonx.ai, the next-generation ...
    May 23, 2023 · We're happy to partner with IBM and to collaborate on the watsonx AI and data platform so that Hugging Face customers can work natively with ...
  66. [66]
    Hugging Face: The Guide for AI Startup Founders - Belitsoft
    Rating 5.0 (1) Mar 30, 2025 · Hugging Face (HF) is a leading open-source platform and community in the machine learning ecosystem. As the “GitHub of Machine Learning”, ...<|control11|><|separator|>
  67. [67]
    The Environmental Impacts of AI -- Primer - Hugging Face
    Sep 3, 2024 · Every aspect of the AI lifecycle consumes natural resources – energy, water, and minerals – and releases greenhouse gases.⚡ Energy ⚡ · 🧩 Missing And Partial... · 🚀 Ways Forward 🚀
  68. [68]
    Model statistics of the 50 most downloaded entities on Hugging Face
    Oct 13, 2025 · The analysis of the 50 most downloaded entities on the Hugging Face Hub (80.22% of total Hub downloads) shows that: Among all open-source models ...
  69. [69]
    Ethics and Bias in AI ‍ ‍ - Hugging Face Community Computer ...
    AI Ethics can be defined as the set of values, principles, and techniques that employ widely accepted standards of right and wrong to guide moral conduct.<|separator|>
  70. [70]
    Training Data Transparency in AI: Tools, Trends, and Policy ...
    Dec 5, 2023 · We are seeing a trend towards decreasing data transparency from developers, especially for models developed for commercial applications of ML.