Experience the Latest Version of Gradio 5 at High Speed! Used by Over 2 Million Users; Selected for ACCV'24, LoLI-Street Low-light Image Enhancement Dataset Launched
Since its launch, Gradio has been used by more than 2 million users per month and plays a key role in the AI development ecosystem. Its concise code and intuitive interface make it easy to transform complex machine learning models into user-friendly web applications while ensuring the security and accessibility of the applications.
The Gradio team recently released the latest version, Gradio 5 stable version, which has made major upgrades in real-time applications and streaming media, with lower latency and a smoother experience.In order to let everyone experience the technical improvements of Gradio 5 earlier and more conveniently, the hyper.ai official website uses Gradio to deploy two popular tutorials, which can be cloned and used with one click:
1. Depth Pro generates 3D depth maps instantly
Run online:https://go.hyper.ai/bSp3h
2. Pyramid Flow generates ultra-high-definition video demo in one minute
Run online:https://go.hyper.ai/njiHn
From October 21st to October 25th, hyper.ai official website updates:
* Selected high-quality tutorials: 2
* High-quality public datasets: 10
* Community article selection: 4 articles
* Popular encyclopedia entries: 5
* Top conferences with deadlines in November: 7
Visit the official website:hyper.ai
Selected Public Tutorials
1. Depth Pro generates 3D depth maps instantly
Depth Pro is an open source zero-shot metric monocular depth estimation (Depth Estimation) basic model that can quickly generate high-resolution 3D depth maps from a single 2D image. This model is not only fast, taking only 0.3 seconds, but also provides metric-level depth information, and the generated depth map has real world scale. The project can generate a front-end interactive interface through the Gradio interface. The relevant models and dependencies have been deployed, and you can experience it by cloning it with one click.
Run online:https://go.hyper.ai/bSp3h
2. Pyramid Flow generates ultra-high-definition video demo in one minute
Pyramid Flow is an open source ultra-high-definition video generation model. This model can generate high-quality videos with a maximum length of 10 seconds, a resolution of up to 1280×768, and a frame rate of 24fps based on text descriptions. Its core technology is the pyramid flow matching algorithm, which decomposes the video generation process into multiple stages with different resolutions to improve generation efficiency and quality. Run the container according to the tutorial and directly copy the API address to generate ultra-high-definition videos.
Run online:https://go.hyper.ai/njiHn
💡We have also established a Stable Diffusion tutorial exchange group. Welcome friends to scan the QR code and remark [SD tutorial] to join the group to discuss various technical issues and share application results~
Selected public datasets
1. LoLI-Street low-light image enhancement dataset
This dataset consists of 33k pairs of low-light and well-exposed images from developed urban street scenes, covering 19k object categories for object detection. It also includes 1k real low-light test images for testing low-light image enhancement (LLIE) models under realistic conditions. This dataset is essential for many computer vision tasks, including object detection, tracking, segmentation, and scene understanding. The related results have been accepted by ACCV'24.
Direct use:https://go.hyper.ai/XD7kV
2. BC-Z Robot Learning Dataset
The dataset supports zero-shot task generalization, which enables robots to perform new manipulation tasks through imitation learning without prior experience. It contains more than 25k different manipulation task scenarios, covering 100 diverse manipulation tasks.
Direct use:https://go.hyper.ai/Lg1GC
3. Chinese Traditional Painting Chinese Traditional Painting Dataset
The dataset contains 1k content images and 100 style images. The content images are mostly real scenes such as Jiangnan mountains, lakes, rivers, bridges, buildings, etc., including not only Chinese scenery, but also beautiful scenery such as the Rhine River, the Alps, Yellowstone, and the Grand Canyon.
Direct use:https://go.hyper.ai/wwZqs
4. OpenMathInstruct-2 Math Instruction Tuning Dataset
The dataset contains 14 million question-answer pairs, which is nearly 8 times larger than the previous largest dataset of its kind. By fine-tuning the Llama-3.1-8B-Base model with OpenMathInstruct-2, its performance on the MATH dataset is improved by 15.9% (from 51.9% to 67.8%) over Llama3.1-8B-Instruct.
Direct use:https://go.hyper.ai/fxskH
5. Omni-MATH Mathematical Reasoning Benchmark Dataset
This dataset contains 4,428 rigorously manually annotated competition-level math problems, covering 33 subfields and more than 10 different difficulty levels, from the Olympiad preparatory level to top Olympiad mathematics competitions such as IMO (International Mathematical Olympiad), IMC (International Mathematical Contest) and Putnam Mathematics Competition.
Direct use:https://go.hyper.ai/tYgfN
6. Reasoning Base 20k reasoning base dataset
This dataset is designed to train reasoning models so that they can think about complex problems like humans and then respond. The dataset includes a variety of questions from different fields (science, coding, mathematics, etc.), each with a detailed chain of ideas (COT) and the correct answer. The goal is to enable the model to learn and improve its reasoning process, identify and correct errors, and provide high-quality, detailed responses.
Direct use:https://go.hyper.ai/ssznB
7. Language-Table Robot language label trajectory dataset
The dataset contains nearly 600K language-labeled trajectories to promote the development of more advanced, capable, natural language-interactive robots. By training on a dataset containing hundreds of thousands of language-annotated trajectories, researchers found that the resulting policy can execute 10 times more instructions than before, which describe end-to-end visual-audio-motor skills in the real world.
Direct use:https://go.hyper.ai/bUPXz
8. BridgeData V2 Large-Scale Robot Learning Dataset
The dataset is designed to facilitate scalable robot learning research and contains more than 60,000 robot trajectories collected in 24 different environments. In order to enhance the generalization ability of robots, researchers collected a large amount of task data in multiple environments with different objects, camera positions, and workspace positioning. Each trajectory is accompanied by natural language instructions corresponding to the robot task.
Direct use:https://go.hyper.ai/eqcYW
9. RT-1 Robot Action Real World Robot Dataset
This dataset is used to train the RT-1 model. The high-level skills covered in the dataset include picking up and placing objects, opening and closing drawers, taking objects out of drawers and putting them in, placing thin objects upright, pushing objects down, pulling napkins, and opening jars, covering more than 700 tasks using a variety of different objects.
Direct use:https://go.hyper.ai/8ySHu
10. MedCalc-Bench medical computing dataset
This dataset contains 10,055 training instances and 1,047 test instances, covering 55 different computational tasks. Each instance includes a patient's note, a question to calculate a specific clinical value, the final answer value, and a step-by-step solution. The purpose of MedCalc-Bench is to improve the linguistic and computational reasoning abilities of LLMs in medical settings.
Direct use:https://go.hyper.ai/5bhzs
For more public datasets, please visit:
Community Articles
Researchers from Tohoku University and MIT have launched a new artificial intelligence tool, GNNOpt, which successfully identified 246 materials with solar energy conversion efficiencies exceeding 32% and 296 quantum materials with high quantum weights, greatly accelerating the discovery of energy and quantum materials. This article is a detailed interpretation and sharing of the research paper.
View the full report:https://go.hyper.ai/3uRDH
The 2024 Open Source AI Forum (AI for Science) will be held on November 2! At that time, researchers from many universities will focus on scientific research fields such as medical health, geographic information science, and spatiotemporal complex systems, and bring in-depth technology popularization and industry development trend analysis from different perspectives of academic research and industry application.
Check out the event details:https://go.hyper.ai/MiQ1O
Zhejiang University and Microsoft Research Asia jointly proposed a new unified medical image pre-training framework UniMedI. It uses diagnostic reports as a common semantic space to create a unified representation for medical images of different modalities, successfully integrating 2D and 3D images, and making better use of complex medical data. This article is a detailed interpretation and sharing of the paper.
View the full report:https://go.hyper.ai/MXYTq
In the face of increasingly urgent climate change, a global and industry-wide carbon reduction campaign is underway. CuspAI, which focuses on using AI to explore carbon capture materials, has taken off, and on June 18 this year, it received a seed round of financing of up to 30 million US dollars (about 217 million RMB), becoming one of the largest seed rounds in Europe that year. What is the charm of this startup? You may be able to find the answer by reading this article.
View the full report:https://go.hyper.ai/nErwd
Popular Encyclopedia Articles
1. Pooling
2. Variational Autoencoder VAE
3. Quantum Neural Network (QNN)
4. Paired t-Test
5. Data Augmentation
Here are hundreds of AI-related terms compiled to help you understand "artificial intelligence" here:
November deadline for the summit
One-stop tracking of top AI academic conferences:https://go.hyper.ai/event
The above is all the content of this week’s editor’s selection. If you have resources that you want to include on the hyper.ai official website, you are also welcome to leave a message or submit an article to tell us!
See you next week!
About HyperAIhyper.ai)
HyperAI Super Neural (hyper.ai) is a leading artificial intelligence and high-performance computing community in China.We are committed to becoming the infrastructure in the field of data science in China and providing rich and high-quality public resources for domestic developers. So far, we have:
* Provide domestic accelerated download nodes for 1300+ public data sets
* Includes 400+ classic and popular online tutorials
* Interpretation of 100+ AI4Science paper cases
* Support 500+ related terms search
* Hosting the first complete Apache TVM Chinese documentation in China
Visit the official website to start your learning journey: