current location:Home > Technical Articles > Technology peripherals > AI
- Direction:
- All web3.0 Backend Development Web Front-end Database Operation and Maintenance Development Tools PHP Framework Daily Programming WeChat Applet Common Problem Other Tech CMS Tutorial Java System Tutorial Computer Tutorials Hardware Tutorial Mobile Tutorial Software Tutorial Mobile Game Tutorial
- Classify:
-
- The latest Chinese large model evaluation is released, and Tencent Hunyuan ranks in the Excellent Leaders Quadrant
- Reporters reported on May 6 that SuperCLUE, an authoritative domestic large model evaluation organization, recently released the "Chinese Large Model Benchmark Test 2024 Fourth Quarter Report". Among them, Tencent Hunyuan large model ranks in the first echelon of domestic large models. It is in a leading position in both basic and scenario applications, and is located in the Excellent Leaders Quadrant. SuperCLUE is the domestic authoritative comprehensive evaluation benchmark for general large models. It was formerly the well-known third-party Chinese language understanding test benchmark CLUE (The Chinese Language Understanding Evaluation). SuperCLUE builds a multi-level and multi-dimensional comprehensive evaluation based on the wide application of general large models in academic, industry and user sides.
- AI 1264 2024-05-06 21:13:32
-
- See through 3D representation and generative models of objects: NUS team proposes X-Ray
- Project homepage: https://tau-yihouxiang.github.io/projects/X-Ray/X-Ray.html Paper address: https://arxiv.org/abs/2404.14329 Code address: https://github.com /tau-yihouxiang/X-Ray Dataset: https://huggingface.co/datasets/yihouxiang/X-Ray Currently, artificial intelligence is developing rapidly in the field of human intelligence. In computer vision, image and video generation technology has become increasingly mature, such as Midjourney
- AI 1052 2024-05-06 18:30:13
-
- 58 lines of code scale Llama 3 to 1 million contexts, any fine-tuned version is applicable
- Llama3, the majestic king of open source, the original context window is only... 8k, which makes me swallow back the words "it smells so good". Today, when 32k is the starting point and 100k is common, is this intentional to leave room for contributions to the open source community? The open source community certainly didn't miss this opportunity: now with just 58 lines of code, any fine-tuned version of Llama370b can automatically scale to 1048k (one million) contexts. Behind the scenes is a LoRA, extracted from a fine-tuned version of Llama370BInstruct that extends good context, and the file is only 800mb. Next, using Mergekit, you can run it with other models of the same architecture or merge it directly into the model. 1048k context used
- AI 1177 2024-05-06 18:10:08
-
- ICLR 2024 Spotlight | Negative label mining facilitates CLIP-based out-of-distribution detection tasks
- As machine learning models are increasingly used in open-world scenarios, how to effectively identify and process out-of-distribution (OOD) data has become an important research area. The presence of out-of-distribution data can lead to model overconfidence and incorrect predictions, which is particularly dangerous in safety-critical applications such as autonomous driving and medical diagnostics. Therefore, developing an effective OOD detection mechanism is crucial to improving the safety and reliability of the model in practical applications. Traditional OOD detection methods mainly focus on a single pattern, especially image data, while ignoring other potentially useful information sources, such as text data. With the rise of visual-language models (VLMs), they have demonstrated powerful capabilities in multi-modal learning scenarios.
- AI 1050 2024-05-06 18:04:24
-
- Regarding artificial intelligence, the secrets experts don't want to reveal!
- Today, artificial intelligence (AI) has become an integral part of our lives. Experts often point to the advantages of artificial intelligence, but may not be quick to mention the problems with the technology. This article digs deep into the hidden secrets that artificial intelligence experts are reluctant to reveal. 1. AI is not foolproof AI systems can learn from large data sets and improve the accuracy of numbers, but they are not perfect. AI models learn from large data sets, and their performance is closely related to the volume and diversity of this data. Bias in the training data can also be translated into the results of artificial intelligence, which can also mimic human-like errors. 2. Data privacy issues Artificial intelligence systems require large amounts of data to operate effectively. Although existing data is often anonymized
- AI 488 2024-05-06 17:55:20
-
- How does RFID technology enhance security measures?
- Radio frequency identification (RFID) technology dates back to the early 1980s, but it has only been widely adopted in recent years. This efficient, cost-effective and innovative technology plays a subtle but powerful role in organizations around the world, with benefits more immediate than the security enhancements it provides. This article will take a closer look at how RFID technology can be used to enhance enterprise security. Strong access security ensures that all organizations should properly protect their facilities to be safe and secure. Failure to do so may result in loss of assets and put employees and customers at risk. With RFID readers, businesses can ensure that only authorized individuals have access to their websites. Think of RFID access cards as a modern solution to keys. With just one card, individuals can easily enter
- AI 439 2024-05-06 17:46:04
-
- New research highlights talent shortages and strategic gaps in GenAI applications
- A recent study shows that U.S. businesses are enthusiastic about the potential of productive forms of artificial intelligence (GenAI) to improve their business and employee productivity. Yet behind the heightened enthusiasm, leaders see gaps in understanding, a lack of strategic planning and a lack of talent as barriers to realizing and measuring the full value of technology. Earlier this year, Coleman Parks Research, with support from SAS, surveyed 300 U.S. GenAI strategy or data analytics decision-makers to explore the key barriers facing investments and organizations in the field. For the study, Coleman Parks also surveyed leaders outside the United States. These global results will be published later this year. The message that can be read in this US executive summary is GenAI Challenges and Potential Demonstration: How to Achieve Competitive Advantage
- AI 996 2024-05-06 17:25:01
-
- Share several .NET open source AI and LLM related project frameworks
- The development of artificial intelligence (AI) technologies is in full swing today, and they have shown great potential and influence in various fields. Today Dayao will share with you 4 .NET open source AI model LLM related project frameworks, hoping to provide you with some reference. https://github.com/YSGStudyHards/DotNetGuide/blob/main/docs/DotNet/DotNetProjectPicks.mdSemanticKernelSemanticKernel is an open source software development kit (SDK) designed to integrate large language models (LLM) such as OpenAI, Azure
- AI 799 2024-05-06 16:43:12
-
- OpenAI Sora makes creators' jobs easy, and they believe they won't be easily replaced
- According to news on May 6, early testers of OpenAI’s text generation video tool Sora seem to be relieved. Rather than panicking about being replaced by this tool, they find it makes their work more interesting. In February this year, artificial intelligence startup OpenAI officially launched Sora, a tool designed to "deeply understand and simulate the changing real world." Sora is known for its unique text-to-video conversion technology that generates feature-length, high-quality videos up to a minute long, and its proof-of-concept video generated widespread attention across the web upon its release. However, for some Hollywood practitioners, the emergence of this technology undoubtedly brings a certain degree of threat. Veteran filmmaker Tyler Perry (
- AI 1157 2024-05-06 16:34:09
-
- As demand grows in the artificial intelligence era, AWS, Microsoft, and Google continue to invest in cloud computing
- The growth of the three cloud computing giants shows no sign of slowing down until 2024, with Amazon, Microsoft, and Google all generating more revenue in cloud computing than ever before. All three cloud vendors have recently reported earnings, continuing their multi-year strategy of consistent revenue growth. On April 25, both Google and Microsoft announced their results. In the first quarter of Alphabet’s fiscal year 2024, Google Cloud’s revenue was US$9.57 billion, a year-on-year increase of 28%. Microsoft's cloud revenue was $35.1 billion, a year-over-year increase of 23%. On April 30, Amazon Web Services (AWS) reported revenue of US$25 billion, a year-on-year increase of 17%, ranking among the three giants. Cloud computing providers have a lot to be happy about, with the growth rates of the three market leaders over the past
- AI 736 2024-05-06 16:22:20
-
- Tesla robots work in factories, Musk: The degree of freedom of hands will reach 22 this year!
- The latest video of Tesla's robot Optimus is released, and it can already work in the factory. At normal speed, it sorts batteries (Tesla's 4680 batteries) like this: The official also released what it looks like at 20x speed - on a small "workstation", picking and picking and picking: This time it is released One of the highlights of the video is that Optimus completes this work in the factory, completely autonomously, without human intervention throughout the process. And from the perspective of Optimus, it can also pick up and place the crooked battery, focusing on automatic error correction: Regarding Optimus's hand, NVIDIA scientist Jim Fan gave a high evaluation: Optimus's hand is the world's five-fingered robot. One of the most dexterous. Its hands are not only tactile
- AI 1204 2024-05-06 16:13:01
-
- LeCun on the moon? Nankai and Byte open source StoryDiffusion to make multi-picture comics and long videos more coherent
- Two days ago, Turing Award winner Yann LeCun reprinted the long comic "Go to the Moon and Explore Yourself", which aroused heated discussion among netizens. In the paper "StoryDiffusion: Consistent Self-Attention for long-range image and video generation", the research team proposed a new method called StoryDiffusion for generating consistent images and videos describing complex scenes. The research on these comics comes from institutions such as Nankai University and ByteDance. Paper address: https://arxiv.org/pdf/2405.01434v1 Project homepage: https:
- AI 953 2024-05-06 16:10:31
-
- For only $250, Hugging Face's technical director teaches you how to fine-tune Llama 3 step by step
- The familiar open source large language models such as Llama3 launched by Meta, Mistral and Mixtral models launched by MistralAI, and Jamba launched by AI21 Lab have become competitors of OpenAI. In most cases, users need to fine-tune these open source models based on their own data to fully unleash the model's potential. It is not difficult to fine-tune a large language model (such as Mistral) compared to a small one using Q-Learning on a single GPU, but efficient fine-tuning of a large model like Llama370b or Mixtral has remained a challenge until now. Therefore, Philipp Sch, technical director of HuggingFace
- AI 1199 2024-05-06 15:52:35
-
- Good news for developers! GitHub launches the AI native development environment Copilot Workspace with 4 core functions. The preview version has been released!
- Compiled | Produced by Yifeng | 51CTO Technology Stack (WeChat ID: blog51cto) Is there a new artifact in the AI programming world? GitHub launched Copilot Workspace, an artificial intelligence native development environment. It aims to allow "any developer to start with an idea, create code, and then proceed to software using natural language." Image developers can use natural language to brainstorm, plan, build, test, and run code faster and easier than ever before. In 2023, GitHub Copilot Workspace attracted user attention for the first time. Now released as a technology preview, developers have shown strong interest in this development tool and have signed up for the waiting list. https://github.co
- AI 869 2024-05-06 15:49:01
-
- Take a look at the current front-end editor
- When helping some students remotely modify the code, I found that the editors used by everyone are various: VSCode, WebStorm and even Sublime... This can't help but remind me: [When I first learned the front-end, I really Tried many different options on the editor]. So far, I have become a "hardcore fan" of using VSCode and have even "forgot" that other editors exist. You can think of this as a kind of "focus." However, this also limits our choices. Therefore, today we will take a look at the current front-end editors. After all, "a hundred schools of thought contend" will "collide" with more inspiration: notepad++ Advantages: Free and open source: Notepad++ is a free
- AI 811 2024-05-06 15:30:16