Enlarge/ Still of Procreate CEO James Cuda from a video posted to X. (credit: Procreate)
On Sunday, Procreate announced that it will not incorporate generative AI into its popular iPad illustration app. The decision comes in response to an ongoing backlash from some parts of the art community, which has raised concerns about the ethical implications and potential consequences of AI use in creative industries.
"Generative AI is ripping the humanity out of things," Procreate wrote on its website. "Built on a foundation of theft, the technology is steering us toward a barren future."
In a video posted on X, Procreate CEO James Cuda laid out his company's stance, saying, "We’re not going to be introducing any generative AI into our products. I don’t like what’s happening to the industry, and I don’t like what it’s doing to artists."
Enlarge/ AI-generated image by FLUX.1 dev: "A beautiful queen of the universe holding up her hands, face in the background." (credit: FLUX.1)
On Thursday, AI-startup Black Forest Labs announced the launch of its company and the release of its first suite of text-to-image AI models, called FLUX.1. The German-based company, founded by researchers who developed the technology behind Stable Diffusion and invented the latent diffusion technique, aims to create advanced generative AI for images and videos.
The launch of FLUX.1 comes about seven weeks after Stability AI's troubled release of Stable Diffusion 3 Medium in mid-June. Stability AI's offering faced widespread criticism among image-synthesis hobbyists for its poor performance in generating human anatomy, with users sharing examples of distorted limbs and bodies across social media. That problematic launch followed the earlier departure of three key engineers from Stability AI—Robin Rombach, Andreas Blattmann, and Dominik Lorenz—who went on to found Black Forest Labs along with latent diffusion co-developer Patrick Esser and others.
Black Forest Labs launched with the release of three FLUX.1 text-to-image models: a high-end commercial "pro" version, a mid-range "dev" version with open weights for non-commercial use, and a faster open-weights "schnell" version ("schnell" means quick or fast in German). Black Forest Labs claims its models outperform existing options like Midjourney and DALL-E in areas such as image quality and adherence to text prompts.
For years, Nvidia has dominated many machine learning benchmarks, and now there are two more notches in its belt.
MLPerf, the AI benchmarking suite sometimes called “the Olympics of machine learning,” has released a new set of training tests to help make more and better apples-to-apples comparisons between competing computer systems. One of MLPerf’s new tests concerns fine-tuning of large language models, a process that takes an existing trained model and trains it a bit more with specialized knowledge to make it fit for a particular purpose. The other is for graph neural networks, a type of machine learning behind some literature databases, fraud detection in financial systems, and social networks.
Even with the additions and the participation of computers using Google’s and Intel’s AI accelerators, systems powered by Nvidia’s Hopper architecture dominated the results once again. One system that included 11,616 Nvidia H100 GPUs—the largest collection yet—topped each of the nine benchmarks, setting records in five of them (including the two new benchmarks).
“If you just throw hardware at the problem, it’s not a given that you’re going to improve.” —Dave Salvator, Nvidia
The 11,616-H100 system is “the biggest we’ve ever done,” says Dave Salvator, director of accelerated computing products at Nvidia. It smashed through the GPT-3 training trial in less than 3.5 minutes. A 512-GPU system, for comparison, took about 51 minutes. (Note that the GPT-3 task is not a full training, which could take weeks and cost millions of dollars. Instead, the computers train on a representative portion of the data, at an agreed-upon point well before completion.)
Compared to Nvidia’s largest entrant on GPT-3 last year, a 3,584 H100 computer, the 3.5-minute result represents a 3.2-fold improvement. You might expect that just from the difference in the size of these systems, but in AI computing that isn’t always the case, explains Salvator. “If you just throw hardware at the problem, it’s not a given that you’re going to improve,” he says.
“We are getting essentially linear scaling,” says Salvator. By that he means that twice as many GPUs lead to a halved training time. “[That] represents a great achievement from our engineering teams,” he adds.
Competitors are also getting closer to linear scaling. This round Intel deployed a system using 1,024 GPUs that performed the GPT-3 task in 67 minutes versus a computer one-fourth the size that took 224 minutes six months ago. Google’s largest GPT-3 entry used 12-times the number of TPU v5p accelerators as its smallest entry and performed its task nine times as fast.
Linear scaling is going to be particularly important for upcoming “AI factories” housing 100,000 GPUs or more, Salvator says. He says to expect one such data center to come online this year, and another, using Nvidia’s next architecture, Blackwell, to startup in 2025.
Nvidia’s streak continues
Nvidia continued to boost training times despite using the same architecture, Hopper, as it did in last year’s training results. That’s all down to software improvements, says Salvator. “Typically, we’ll get a 2-2.5x [boost] from software after a new architecture is released,” he says.
For GPT-3 training, Nvidia logged a 27 percent improvement from the June 2023 MLPerf benchmarks. Salvator says there were several software changes behind the boost. For example, Nvidia engineers tuned up Hopper’s use of less accurate, 8-bit floating point operations by trimming unnecessary conversions between 8-bit and 16-bit numbers and better targeting of which layers of a neural network could use the lower precision number format. They also found a more intelligent way to adjust the power budget of each chip’s compute engines, and sped communication among GPUs in a way that Salvator likened to “buttering your toast while it’s still in the toaster.”
Additionally, the company implemented a scheme called flash attention. Invented in the Stanford University laboratory of Samba Nova founder Chris Re, flash attention is an algorithm that speeds transformer networks by minimizing writes to memory. When it first showed up in MLPerf benchmarks, flash attention shaved as much as 10 percent from training times. (Intel, too, used a version of flash attention but not for GPT-3. It instead used the algorithm for one of the new benchmarks, fine-tuning.)
Using other software and network tricks, Nvidia delivered an 80 percent speedup in the text-to-image test, Stable Diffusion, versus its submission in November 2023.
New benchmarks
MLPerf adds new benchmarks and upgrades old ones to stay relevant to what’s happening in the AI industry. This year saw the addition of fine-tuning and graph neural networks.
Fine tuning takes an already trained LLM and specializes it for use in a particular field. Nvidia, for example took a trained 43-billion-parameter model and trained it on the GPU-maker’s design files and documentation to create ChipNeMo, an AI intended to boost the productivity of its chip designers. At the time, the company’s chief technology officer Bill Dally said that training an LLM was like giving it a liberal arts education, and fine tuning was like sending it to graduate school.
The MLPerf benchmark takes a pretrained Llama-2-70B model and asks the system to fine tune it using a dataset of government documents with the goal of generating more accurate document summaries.
There are several ways to do fine-tuning. MLPerf chose one called low-rank adaptation (LoRA). The method winds up training only a small portion of the LLM’s parameters leading to a 3-fold lower burden on hardware and reduced use of memory and storage versus other methods, according to the organization.
The other new benchmark involved a graph neural network (GNN). These are for problems that can be represented by a very large set of interconnected nodes, such as a social network or a recommender system. Compared to other AI tasks, GNNs require a lot of communication between nodes in a computer.
The benchmark trained a GNN on a database that shows relationships about academic authors, papers, and institutes—a graph with 547 million nodes and 5.8 billion edges. The neural network was then trained to predict the right label for each node in the graph.
Future fights
Training rounds in 2025 may see head-to-head contests comparing new accelerators from AMD, Intel, and Nvidia. AMD’s MI300 series was launched about six months ago, and a memory-boosted upgrade the MI325x is planned for the end of 2024, with the next generation MI350 slated for 2025. Intel says its Gaudi 3, generally available to computer makers later this year, will appear in MLPerf’s upcoming inferencing benchmarks. Intel executives have said the new chip has the capacity to beat H100 at training LLMs. But the victory may be short-lived, as Nvidia has unveiled a new architecture, Blackwell, which is planned for late this year.
Salad, a cloud computing and AI tech company, is renting high-end graphics cards found in gamers’ computers and using all that power to create AI-generated pornography. In return, the gamers who lend the company their GPUs are paid in Fortnite skins, Minecraft cosmetics, Roblox bux, and other gaming-related gift cards…
The day after The New York Times sued OpenAI for copyright infringement, the author and systems architect Daniel Jeffries wrote an essay-length tweet arguing that the Times “has a near zero probability of winning” its lawsuit. As we write this, it has been retweeted 288 times and received 885,000 views.
“Trying to get everyone to license training data is not going to work because that's not what copyright is about,” Jeffries wrote. “Copyright law is about preventing people from producing exact copies or near exact copies of content and posting it for commercial gain. Period. Anyone who tells you otherwise is lying or simply does not understand how copyright works.”
This article is written by two authors. One of us is a journalist who has been on the copyright beat for nearly 20 years. The other is a law professor who has taught dozens of courses on IP and Internet law. We’re pretty sure we understand how copyright works. And we’re here to warn the AI community that it needs to take these lawsuits seriously.
On Friday, Bloomberg reported that Reddit has signed a contract allowing an unnamed AI company to train its models on the site's content, according to people familiar with the matter. The move comes as the social media platform nears the introduction of its initial public offering (IPO), which could happen as soon as next month.
Reddit initially revealed the deal, which is reported to be worth $60 million a year, earlier in 2024 to potential investors of an anticipated IPO, Bloomberg said. The Bloomberg source speculates that the contract could serve as a model for future agreements with other AI companies.
After an era where AI companies utilized AI training data without expressly seeking any rightsholder permission, some tech firms have more recently begun entering deals where some content used for training AI models similar to GPT-4 (which runs the paid version of ChatGPT) comes under license. In December, for example, OpenAI signed an agreement with German publisher Axel Springer (publisher of Politico and Business Insider) for access to its articles. Previously, OpenAI has struck deals with other organizations, including the Associated Press. Reportedly, OpenAI is also in licensing talks with CNN, Fox, and Time, among others.
Přihlásit se
