Open weights AI

google-and-meta-update-their-ai-models-amid-the-rise-of-“alphachip”

Google and Meta update their AI models amid the rise of “AlphaChip”

AI, AlphaChip, Biz & IT, Gemini, Gemini 1.5, Google, Google Gemini, Llama 3.2, machine learning, Meta, Meta Llama, Open weights AI, openai, Runway / /

Running the AI News Gauntlet —

News about Gemini updates, Llama 3.2, and Google’s new AI-powered chip designer.

–

Cyberpunk concept showing a man running along a futuristic path full of monitors.

Enlarge / There’s been a lot of AI news this week, and covering it sometimes feels like running through a hall full of danging CRTs, just like this Getty Images illustration.

It’s been a wildly busy week in AI news thanks to OpenAI, including a controversial blog post from CEO Sam Altman, the wide rollout of Advanced Voice Mode, 5GW data center rumors, major staff shake-ups, and dramatic restructuring plans.

But the rest of the AI world doesn’t march to the same beat, doing its own thing and churning out new AI models and research by the minute. Here’s a roundup of some other notable AI news from the past week.

Google Gemini updates

On Tuesday, Google announced updates to its Gemini model lineup, including the release of two new production-ready models that iterate on past releases: Gemini-1.5-Pro-002 and Gemini-1.5-Flash-002. The company reported improvements in overall quality, with notable gains in math, long context handling, and vision tasks. Google claims a 7 percent increase in performance on the MMLU-Pro benchmark and a 20 percent improvement in math-related tasks. But as you know, if you’ve been reading Ars Technica for a while, AI typically benchmarks aren’t as useful as we would like them to be.

Along with model upgrades, Google introduced substantial price reductions for Gemini 1.5 Pro, cutting input token costs by 64 percent and output token costs by 52 percent for prompts under 128,000 tokens. As AI researcher Simon Willison noted on his blog, “For comparison, GPT-4o is currently $5/[million tokens] input and $15/m output and Claude 3.5 Sonnet is $3/m input and $15/m output. Gemini 1.5 Pro was already the cheapest of the frontier models and now it’s even cheaper.”

Google also increased rate limits, with Gemini 1.5 Flash now supporting 2,000 requests per minute and Gemini 1.5 Pro handling 1,000 requests per minute. Google reports that the latest models offer twice the output speed and three times lower latency compared to previous versions. These changes may make it easier and more cost-effective for developers to build applications with Gemini than before.

Meta launches Llama 3.2

On Wednesday, Meta announced the release of Llama 3.2, a significant update to its open-weights AI model lineup that we have covered extensively in the past. The new release includes vision-capable large language models (LLMs) in 11 billion and 90B parameter sizes, as well as lightweight text-only models of 1B and 3B parameters designed for edge and mobile devices. Meta claims the vision models are competitive with leading closed-source models on image recognition and visual understanding tasks, while the smaller models reportedly outperform similar-sized competitors on various text-based tasks.

Willison did some experiments with some of the smaller 3.2 models and reported impressive results for the models’ size. AI researcher Ethan Mollick showed off running Llama 3.2 on his iPhone using an app called PocketPal.

Meta also introduced the first official “Llama Stack” distributions, created to simplify development and deployment across different environments. As with previous releases, Meta is making the models available for free download, with license restrictions. The new models support long context windows of up to 128,000 tokens.

Google’s AlphaChip AI speeds up chip design

On Thursday, Google DeepMind announced what appears to be a significant advancement in AI-driven electronic chip design, AlphaChip. It began as a research project in 2020 and is now a reinforcement learning method for designing chip layouts. Google has reportedly used AlphaChip to create “superhuman chip layouts” in the last three generations of its Tensor Processing Units (TPUs), which are chips similar to GPUs designed to accelerate AI operations. Google claims AlphaChip can generate high-quality chip layouts in hours, compared to weeks or months of human effort. (Reportedly, Nvidia has also been using AI to help design its chips.)

Notably, Google also released a pre-trained checkpoint of AlphaChip on GitHub, sharing the model weights with the public. The company reported that AlphaChip’s impact has already extended beyond Google, with chip design companies like MediaTek adopting and building on the technology for their chips. According to Google, AlphaChip has sparked a new line of research in AI for chip design, potentially optimizing every stage of the chip design cycle from computer architecture to manufacturing.

That wasn’t everything that happened, but those are some major highlights. With the AI industry showing no signs of slowing down at the moment, we’ll see how next week goes.

Google and Meta update their AI models amid the rise of “AlphaChip” Read More »

exponential-growth-brews-1-million-ai-models-on-hugging-face

Exponential growth brews 1 million AI models on Hugging Face

AI, AI host, AI models, AI platform, Biz & IT, Clément Delangue, Hugging Face, machine learning, open source AI, Open weights AI / /

The sand has come alive —

Hugging Face cites community-driven customization as fuel for diverse AI model boom.

–

The Hugging Face logo in front of shipping containers.

On Thursday, AI hosting platform Hugging Face surpassed 1 million AI model listings for the first time, marking a milestone in the rapidly expanding field of machine learning. An AI model is a computer program (often using a neural network) trained on data to perform specific tasks or make predictions. The platform, which started as a chatbot app in 2016 before pivoting to become an open source hub for AI models in 2020, now hosts a wide array of tools for developers and researchers.

The machine-learning field represents a far bigger world than just large language models (LLMs) like the kind that power ChatGPT. In a post on X, Hugging Face CEO Clément Delangue wrote about how his company hosts many high-profile AI models, like “Llama, Gemma, Phi, Flux, Mistral, Starcoder, Qwen, Stable diffusion, Grok, Whisper, Olmo, Command, Zephyr, OpenELM, Jamba, Yi,” but also “999,984 others.”

The reason why, Delangue says, stems from customization. “Contrary to the ‘1 model to rule them all’ fallacy,” he wrote, “smaller specialized customized optimized models for your use-case, your domain, your language, your hardware and generally your constraints are better. As a matter of fact, something that few people realize is that there are almost as many models on Hugging Face that are private only to one organization—for companies to build AI privately, specifically for their use-cases.”

A Hugging Face-supplied chart showing the number of AI models added to Hugging Face over time, month to month.

Enlarge / A Hugging Face-supplied chart showing the number of AI models added to Hugging Face over time, month to month.

Hugging Face’s transformation into a major AI platform follows the accelerating pace of AI research and development across the tech industry. In just a few years, the number of models hosted on the site has grown dramatically along with interest in the field. On X, Hugging Face product engineer Caleb Fahlgren posted a chart of models created each month on the platform (and a link to other charts), saying, “Models are going exponential month over month and September isn’t even over yet.”

The power of fine-tuning

As hinted by Delangue above, the sheer number of models on the platform stems from the collaborative nature of the platform and the practice of fine-tuning existing models for specific tasks. Fine-tuning means taking an existing model and giving it additional training to add new concepts to its neural network and alter how it produces outputs. Developers and researchers from around the world contribute their results, leading to a large ecosystem.

For example, the platform hosts many variations of Meta’s open-weights Llama models that represent different fine-tuned versions of the original base models, each optimized for specific applications.

Hugging Face’s repository includes models for a wide range of tasks. Browsing its models page shows categories such as image-to-text, visual question answering, and document question answering under the “Multimodal” section. In the “Computer Vision” category, there are sub-categories for depth estimation, object detection, and image generation, among others. Natural language processing tasks like text classification and question answering are also represented, along with audio, tabular, and reinforcement learning (RL) models.

A screenshot of the Hugging Face models page captured on September 26, 2024.

Enlarge / A screenshot of the Hugging Face models page captured on September 26, 2024.

Hugging Face

When sorted for “most downloads,” the Hugging Face models list reveals trends about which AI models people find most useful. At the top, with a massive lead at 163 million downloads, is Audio Spectrogram Transformer from MIT, which classifies audio content like speech, music, and environmental sounds. Following that, with 54.2 million downloads, is BERT from Google, an AI language model that learns to understand English by predicting masked words and sentence relationships, enabling it to assist with various language tasks.

Rounding out the top five AI models are all-MiniLM-L6-v2 (which maps sentences and paragraphs to 384-dimensional dense vector representations, useful for semantic search), Vision Transformer (which processes images as sequences of patches to perform image classification), and OpenAI’s CLIP (which connects images and text, allowing it to classify or describe visual content using natural language).

No matter what the model or the task, the platform just keeps growing. “Today a new repository (model, dataset or space) is created every 10 seconds on HF,” wrote Delangue. “Ultimately, there’s going to be as many models as code repositories and we’ll be here for it!”

Exponential growth brews 1 million AI models on Hugging Face Read More »

debate-over-“open-source-ai”-term-brings-new-push-to-formalize-definition

Debate over “open source AI” term brings new push to formalize definition

AI, AI regulation, AI workshops, All Things Open, Biz & IT, chatgpt, chatgtp, flux, free software, GPT-4o, machine learning, Meta, Open Source, open source AI, Open Source Initiative, open weights, Open weights AI, openai, OSI, Raleigh, SB-1047, source available / /
A man peers over a glass partition, seeking transparency.

Enlarge / A man peers over a glass partition, seeking transparency.

The Open Source Initiative (OSI) recently unveiled its latest draft definition for “open source AI,” aiming to clarify the ambiguous use of the term in the fast-moving field. The move comes as some companies like Meta release trained AI language model weights and code with usage restrictions while using the “open source” label. This has sparked intense debates among free-software advocates about what truly constitutes “open source” in the context of AI.

For instance, Meta’s Llama 3 model, while freely available, doesn’t meet the traditional open source criteria as defined by the OSI for software because it imposes license restrictions on usage due to company size or what type of content is produced with the model. The AI image generator Flux is another “open” model that is not truly open source. Because of this type of ambiguity, we’ve typically described AI models that include code or weights with restrictions or lack accompanying training data with alternative terms like “open-weights” or “source-available.”

To address the issue formally, the OSI—which is well-known for its advocacy for open software standards—has assembled a group of about 70 participants, including researchers, lawyers, policymakers, and activists. Representatives from major tech companies like Meta, Google, and Amazon also joined the effort. The group’s current draft (version 0.0.9) definition of open source AI emphasizes “four fundamental freedoms” reminiscent of those defining free software: giving users of the AI system permission to use it for any purpose without permission, study how it works, modify it for any purpose, and share with or without modifications.

By establishing clear criteria for open source AI, the organization hopes to provide a benchmark against which AI systems can be evaluated. This will likely help developers, researchers, and users make more informed decisions about the AI tools they create, study, or use.

Truly open source AI may also shed light on potential software vulnerabilities of AI systems, since researchers will be able to see how the AI models work behind the scenes. Compare this approach with an opaque system such as OpenAI’s ChatGPT, which is more than just a GPT-4o large language model with a fancy interface—it’s a proprietary system of interlocking models and filters, and its precise architecture is a closely guarded secret.

OSI’s project timeline indicates that a stable version of the “open source AI” definition is expected to be announced in October at the All Things Open 2024 event in Raleigh, North Carolina.

“Permissionless innovation”

In a press release from May, the OSI emphasized the importance of defining what open source AI really means. “AI is different from regular software and forces all stakeholders to review how the Open Source principles apply to this space,” said Stefano Maffulli, executive director of the OSI. “OSI believes that everybody deserves to maintain agency and control of the technology. We also recognize that markets flourish when clear definitions promote transparency, collaboration and permissionless innovation.”

The organization’s most recent draft definition extends beyond just the AI model or its weights, encompassing the entire system and its components.

For an AI system to qualify as open source, it must provide access to what the OSI calls the “preferred form to make modifications.” This includes detailed information about the training data, the full source code used for training and running the system, and the model weights and parameters. All these elements must be available under OSI-approved licenses or terms.

Notably, the draft doesn’t mandate the release of raw training data. Instead, it requires “data information”—detailed metadata about the training data and methods. This includes information on data sources, selection criteria, preprocessing techniques, and other relevant details that would allow a skilled person to re-create a similar system.

The “data information” approach aims to provide transparency and replicability without necessarily disclosing the actual dataset, ostensibly addressing potential privacy or copyright concerns while sticking to open source principles, though that particular point may be up for further debate.

“The most interesting thing about [the definition] is that they’re allowing training data to NOT be released,” said independent AI researcher Simon Willison in a brief Ars interview about the OSI’s proposal. “It’s an eminently pragmatic approach—if they didn’t allow that, there would be hardly any capable ‘open source’ models.”

Debate over “open source AI” term brings new push to formalize definition Read More »

google-goes-“open-ai”-with-gemma,-a-free,-open-weights-chatbot-family

Google goes “open AI” with Gemma, a free, open-weights chatbot family

AI, Biz & IT, chatgpt, chatgtp, Gemini, Google, Google Gemini, LLaMA, Llama 2, machine learning, Meta, Meta Llama, Open AI models, Open weights AI, openai, source-available AI / /

Free hallucinations for all —

Gemma chatbots can run locally, and they reportedly outperform Meta’s Llama 2.

–

The Google Gemma logo

On Wednesday, Google announced a new family of AI language models called Gemma, which are free, open-weights models built on technology similar to the more powerful but closed Gemini models. Unlike Gemini, Gemma models can run locally on a desktop or laptop computer. It’s Google’s first significant open large language model (LLM) release since OpenAI’s ChatGPT started a frenzy for AI chatbots in 2022.

Gemma models come in two sizes: Gemma 2B (2 billion parameters) and Gemma 7B (7 billion parameters), each available in pre-trained and instruction-tuned variants. In AI, parameters are values in a neural network that determine AI model behavior, and weights are a subset of these parameters stored in a file.

Developed by Google DeepMind and other Google AI teams, Gemma pulls from techniques learned during the development of Gemini, which is the family name for Google’s most capable (public-facing) commercial LLMs, including the ones that power its Gemini AI assistant. Google says the name comes from the Latin gemma, which means “precious stone.”

While Gemma is Google’s first major open LLM since the launch of ChatGPT (it has released smaller research models such as FLAN-T5 in the past), it’s not Google’s first contribution to open AI research. The company cites the development of the Transformer architecture, as well as releases like TensorFlow, BERT, T5, and JAX as key contributions, and it would not be controversial to say that those have been important to the field.

A chart of Gemma performance provided by Google. Google says that Gemma outperforms Meta's Llama 2 on several benchmarks.

Enlarge / A chart of Gemma performance provided by Google. Google says that Gemma outperforms Meta’s Llama 2 on several benchmarks.

Owing to lesser capability and high confabulation rates, smaller open-weights LLMs have been more like tech demos until recently, as some larger ones have begun to match GPT-3.5 performance levels. Still, experts see source-available and open-weights AI models as essential steps in ensuring transparency and privacy in chatbots. Google Gemma is not “open source” however, since that term usually refers to a specific type of software license with few restrictions attached.

In reality, Gemma feels like a conspicuous play to match Meta, which has made a big deal out of releasing open-weights models (such as LLaMA and Llama 2) since February of last year. That technique stands in opposition to AI models like OpenAI’s GPT-4 Turbo, which is only available through the ChatGPT application and a cloud API and cannot be run locally. A Reuters report on Gemma focuses on the Meta angle and surmises that Google hopes to attract more developers to its Vertex AI cloud platform.

We have not used Gemma yet; however, Google claims the 7B model outperforms Meta’s Llama 2 7B and 13B models on several benchmarks for math, Python code generation, general knowledge, and commonsense reasoning tasks. It’s available today through Kaggle, a machine-learning community platform, and Hugging Face.

In other news, Google paired the Gemma release with a “Responsible Generative AI Toolkit,” which Google hopes will offer guidance and tools for developing what the company calls “safe and responsible” AI applications.

Google goes “open AI” with Gemma, a free, open-weights chatbot family Read More »