📋 星标文章列表

After a period of uncertainty, confidence is returning to Phoenix’s commercial real estate market. According to Arizona State University’s Center for Real Estate and Finance’s latest Commercial Broker Sentiment Index, or CBSI, the Valley posted a reading of 62.7 — signaling the strongest confidence level among brokers since interest rates began rising. The index, which serves as a forward-looking indicator of market health, suggests that Phoenix’s real estate professionals are seeing brighter prospects ahead and expect continued improvement into 2026. Developed from more than a decade of broker survey data, the CBSI measures how local industry experts view the next six months in terms of leasing, sales and overall investment activity. A score above 50 signals optimism, while anything below 50 indicates contraction. In October 2022, the CBSI was just above 30. With this latest reading at 62.7, brokers are feeling moderately optimistic about current conditions and confident that the city’s commercial real estate market is growing. To explore what’s driving market activity, which sectors are leading the charge and how brokers are adapting to shifting economic realities, ASU News spoke with Tom Johnston, executive director at the Center for Real Estate and Finance, which is housed in the W. P. Carey School of Business. Johnston explains what the data tells us about the market’s trajectory and how the newly expanded ASU center will work with the local real estate community. Related story W. P. Carey School of Business celebrates new center and degree program for real estate, a vital industry for the state Note: Answers have been edited for length and/or clarity. Question: What’s the mission of ASU’s Center for Real Estate and Finance, and what kind of impact do you hope it will have on Arizona? Answer: The center serves as a bridge between ASU, its students and the broader real estate community. My background is in brokerage and operations, but the center focuses on the ...
(全文较长，仅显示前2000字符，查看完整内容)

Access Denied You don't have permission to access "http://www.esri.com/about/newsroom/publications/wherenext/nexttech-generative-ai-retail" on this server. Reference #18.f793517.1763537064.47b155a6 https://errors.edgesuite.net/18.f793517.1763537064.47b155a6

Cat owners know the struggle intimately: fur on every surface, hairballs appearing at the worst moments, and the daily battle of brushing a cat who’d rather do anything else than sit still for grooming sessions. Most grooming tools are either stressful for cats or a constant hassle for humans, and the result is often a home that never quite feels clean, no matter how much you vacuum or sweep throughout the week. PawSwing Neo offers a different way to keep cats happy and homes fur-free without the usual stress, resistance, or effort from either party. With its bionic cat-tongue grooming system, cozy felt house design, and zero electricity required for operation, it turns grooming into a natural, stress-free part of daily life. No batteries, no motors, no fighting with your cat over a brush they absolutely hate—just feline instinct and clever design working together seamlessly. Designer: Andrew Tian Click Here to Buy Now: $179 $289 (39% off). Hurry, only a few left! Real user feedback from thousands of cat households across more than 48 countries highlights PawSwing’s mind-blowing effectiveness in collecting substantial amounts of fur in just one week. Some owners even describe opening the collection boxes to find surprising amounts of shed hair they didn’t even realize their cats were losing daily. The visible results speak for themselves, with up to 80 percent less fur accumulating on furniture, clothes, and floors throughout the home. The Neo’s compact, cube-shaped house is crafted from renewable PET felt that’s durable, scratch-resistant, and soft to the touch for comfortable lounging throughout the day. Its lighter weight at 4.8 kilograms and smaller footprint compared to the Pro version make it easy to move between rooms and fit into any living space without dominating valuable floor area or clashing with your carefully chosen furniture and decor. The adjustable entrance adapts to cats of all sizes, from petite British Shorthairs to fluffy Maine Coons weighing u...
(全文较长，仅显示前2000字符，查看完整内容)

Boy Scouts, how far would you go to earn your Eagle Scout rank? How about traveling to Texas from Spain or South Korea? BSA Troop 1899 Girls at the First Unitarian Church welcomed teens from those countries for the 2024-25 school year and celebrated with them a few months ago as they earned Boy Scouting’s highest rank. Exchange students María Hernandez Cazorla and Seoyoung Kee stayed with U.S. guardians Pierre and Dorothy Krouse, of University Park, while spending a school year at Highland Park High School, María’s sophomore and Seoyoung’s junior years. “Seoyoung actually also earned the highest rank in Korea, which is the Tiger Scout,” Dorothy Krouse said. About the Eagle Scouts: María Hernandez Cazorla is the daughter of Alberto Hernandez Bel and Sonia Cazorla Marin, of Spain. Her Eagle project: María built a large double-sided outdoor blackboard for the new Butterfly Garden at the Dallas Zoo. Seoyoung Kee is the daughter of Hyunjin Jung and Hoigyu Kee, of South Korea. Her Eagle project: Seoyoung built a large double-sided outdoor blackboard for the Wild Earth Preschool area at the Dallas Zoo.

City Council approved a resolution authorizing the city to apply for $1.08 million in state grant funding to expand its accessory dwelling unit (ADU) program aimed at assisting low-income homeowners and tenants. The grant application, submitted through the California Department of Housing and Community Development’s CalHome program, would help finance the construction or rehabilitation of ADUs and junior ADUs (JADUs), also known as “granny flats” or backyard homes. The funding round allows cities to request up to $5 million per application. The City launched its 2nd Unit ADU Program in January with existing state funds totaling more than $5 million from CalHome and the Permanent Local Housing Allocation (PLHA). The initiative provides low-interest loans of up to $225,000 for homeowners to build or convert space into an ADU, with the requirement they rent the units to Pasadena residents on rental assistance programs such as Section 8 for at least seven years. The application period was extended following the Eaton Fire, and out of 18 applicants, just three were deemed fully eligible. Of those, two are currently moving forward with loan processing. The new funding would enable the city to construct an estimated four additional ADUs, bringing the total to as many as 25 units over the next three years. Altogether, Pasadena’s ADU program would receive more than $6.2 million in state support. The program supports the City Council’s strategic goals including improving public facilities, mobility, and housing quality. The Housing Department’s proposal is part of Monday’s consent calendar.

Ongoing discussion about amending language in the Mount Pleasant Code of Ordinances saw local lawmakers recently weigh in on proposed Accessory Dwelling Unit (ADU) standards that could extend beyond the Old Village Historic District. A Nov. 12 first-reading presentation by Mount Pleasant Planning Director Michele Reed outlined planning commission and planning committee recommendations to modify three sections of the code of ordinances. If adopted, the first section would lower the maximum height of principal structures in the Old Village and Old Mount Pleasant. It would also increase the setbacks for a detached accessory structure from six to 10 feet.

Marlington High School honors veterans with moving assembly, tributes to fallen alumni Students, staff and community members gathered for a powerful Veterans Day program featuring music, recognition of service, student-led fundraising and a solemn remembrance of Marlington graduates who gave their lives for their country. THE LATEST - Marlington board celebrates district awards, major donations and program growth - Marlington High School honors veterans with moving assembly, tributes to fallen alumni - Hartville mayor honors veterans; council approves legislation, announces Fill a Cruiser event - Marlington bowling teams eye progress with deep boys roster, rebuilding girls squad - Marlboro trustees certify lighting rates after police levy passes 2-1 - Stark Parks celebrates season with holiday events, hikes and winter adventures - Week 13 H.S. football playoff scoreboard - ConnectWell CIE launches in Stark County to link health and social services - Christkindl Markt ushers in holiday spirit at Canton Museum of Art - Week 13 H.S. football playoff preview: Games and players to watch

On behalf of the Village of Southampton Commission on Veterans Patriotic Events, I would like to express my sincere gratitude to the following individuals and groups who helped make this year’s Veterans Day celebration a success: Our veterans, volunteers and donors; Emily Paccasassi, for singing the national anthem; the Southampton High School Chorus and Band and James Tartaglia, its director; our guest speaker, Vietnam veteran nurse Joan Furey, who earned a Bronze Star for her service in Vietnam in 1968; the Southampton Village Police Department, EMTs, the Village Department of Public Works, and the fire department for their presence and service in Agawam Park; the Boy Scouts, Cub Scouts and the 106th Civil Air Patrol. I would also like to thank Sean King from Sean’s Place, for his generous donation of 100 servings of assorted homemade soups; Keith Davis from Golden Pear, for his donation of bread and butter; Liz Tureski-Blake, for her homemade salads; and Stop & Shop, for their cake and cupcakes, which fed our veterans and volunteers at our annual luncheon. And to all the amazing group of volunteer women who always show up to honor our veterans, and who make this day amazing with great food and service. Thank you. Thank you. Thank you. In remembrance and gratitude for my grandfather William August Frankenbach, and to my parents, William and Colleen Frankenbach, to carry on their legacy in Southampton. Lisa Frankenbach McWilliams Southampton The Board of Fire Commissioners of the Hampton Bays Fire District, and the chiefs of the Hampton Bays Fire Department, have been made aware of recent reports and social media posts concerning the reported presence of Immigration and Customs Enforcement... Thank you to everyone who read my letter last week [“Baffling Decision,” Letters, November 13], and a special thank you to Stephen Ring for circulating it on his famous listserv. It was an unexpected honor, and I’m grateful for the... “We the People … All men are created equal...
(全文较长，仅显示前2000字符，查看完整内容)

AIR (AI Realtor), a Dubai-founded proptech company, has launched what it describes as the world’s first fully AI-native real estate brokerage platform, marking a major step toward the UAE’s ambition to become a global leader in artificial intelligence. Developed within the framework of the National Strategy for Artificial Intelligence 2031, AIR is a homegrown innovation built entirely in Dubai’s digital ecosystem. The platform aims to transform property transactions by combining human insight with advanced AI to create simpler, faster and more transparent real estate experiences. “No human can monitor every listing, price movement, and buyer signal across the entire city, but AIR can, instantly,” said Milad Monshipour, Founder and CEO of AIR. “Our mission is to become the most trusted advisor in the country, in the region, and soon globally – transforming how people buy and sell properties, making it fundamentally simpler, smarter, and more trustworthy through human-centred AI.” At the heart of the platform is AIR Brain, a self-learning engine trained on Dubai’s property data. It integrates a suite of intelligent tools, including AIR GPT, a comprehensive real estate knowledge model; AIR Match, a personalised property search tool; AIR Value, a live valuation engine; and AIR Assistant, which automates scheduling and CRM tasks. The platform is designed to empower agents, streamline operations and deliver data-driven transparency. By automating routine tasks, AIR enables brokers to focus on client relationships and negotiations, while giving buyers and sellers access to real-time, explainable insights. “AIR doesn’t evolve real estate, it rewrites it,” Monshipour said. “For the first time, AI understands Dubai’s property market in real time, empowering agents to move faster, act smarter, and redefine what trust and performance mean in this industry.” Backed by the UAE’s strong digital infrastructure, regulatory support and commitment to open data, AIR reflects the countr...
(全文较长，仅显示前2000字符，查看完整内容)

In this blog post we'll take a look at what it takes to build the technology behind GitHub CoPilot, an application that provides suggestions to programmers as they code. In this step by step guide, we'll learn how to train a large GPT-2 model called CodeParrot 🦜, entirely from scratch. CodeParrot can auto-complete your Python code - give it a spin here. Let's get to building it from scratch! The first thing we need is a large training dataset. With the goal to train a Python code generation model, we accessed the GitHub dump available on Google's BigQuery and filtered for all Python files. The result is a 180 GB dataset with 20 million files (available here). After initial training experiments, we found that the duplicates in the dataset severely impacted the model performance. Further investigating the dataset we found that: - 0.1% of the unique files make up 15% of all files - 1% of the unique files make up 35% of all files - 10% of the unique files make up 66% of all files You can learn more about our findings in this Twitter thread. We removed the duplicates and applied the same cleaning heuristics found in the Codex paper. Codex is the model behind CoPilot and is a GPT-3 model fine-tuned on GitHub code. The cleaned dataset is still 50GB big and available on the Hugging Face Hub: codeparrot-clean. With that we can setup a new tokenizer and train a model. First we need a tokenizer. Let's train one specifically on code so it splits code tokens well. We can take an existing tokenizer (e.g. GPT-2) and directly train it on our own dataset with the train_new_from_iterator() method. We then push it to the Hub. Note that we omit imports, arguments parsing and logging from the code examples to keep the code blocks compact. But you'll find the full code including preprocessing and downstream task evaluation here. # Iterator for Training def batch_iterator(batch_size=10): for _ in tqdm(range(0, args.n_examples, batch_size)): yield [next(iter_dataset)["content"] for _ in ra...
(全文较长，仅显示前2000字符，查看完整内容)

We're excited to share our first custom Deep Reinforcement Learning environment: Snowball Fight 1vs1 🎉. Snowball Fight is a game made with Unity ML-Agents, where you shoot snowballs against a Deep Reinforcement Learning agent. The game is hosted on Hugging Face Spaces. In this post, we'll cover the ecosystem we are working on for Deep Reinforcement Learning researchers and enthusiasts that use Unity ML-Agents. The Unity Machine Learning Agents Toolkit is an open source library that allows you to build games and simulations with Unity game engine to serve as environments for training intelligent agents. With this first step, our goal is to build an ecosystem on Hugging Face for Deep Reinforcement Learning researchers and enthusiasts that uses ML-Agents, with three features. - Building and sharing custom environments. We are developing and sharing exciting environments to experiment with new problems: snowball fights, racing, puzzles... All of them will be open source and hosted on the Hugging Face's Hub. - Allowing you to easily host your environments, save models and share them on the Hugging Face Hub. We have already published the Snowball Fight training environment here, but there will be more to come! - You can now easily host your demos on Spaces and showcase your results quickly with the rest of the ecosystem. If you're using ML-Agents or interested in Deep Reinforcement Learning and want to be part of the conversation, you can join our discord server. We just added two channels (and we'll add more in the future): - Deep Reinforcement Learning - ML-Agents Our discord is the place where you can exchange about Hugging Face, NLP, Deep RL, and more! It's also in this discord that we'll announce all our new environments and features in the future. In the coming weeks and months, we will be extending the ecosystem by: - Writing some technical tutorials on ML-Agents. - Working on a Snowball Fight 2vs2 version, where the agents will collaborate in teams using MA-POCA, ...
(全文较长，仅显示前2000字符，查看完整内容)

Chippewa Falls Boy Scouts cook up another “Scouts Giving” meal Recommended articles "Jeffrey definitely had dirt on Trump," Mark alleged in an interview with NewsNation ABC and ESPN have made a lot of college football fans famous over the years. From the infamous college football superfans, like Buck-I-Guy and Big Nut at Ohio State, to the co-eds like Jenn Sterger and Katherine Webb, several college football fans have gone viral thanks to ABC and ESPN's camera ... A female Georgia fan has the internet buzzing after she was shown on the ESPN broadcast during Saturday's 35-10 win over Texas. (WFRV) – The Shawano County man charged with homicide had his bond set at $1 million as the details of the incident were released. According to a criminal complaint obtained by Local 5, 74-year-old Brent Hofman was officially charged with homicide and attempted homicide. The criminal complaint provided details of the incident. On November 12 […] President Trump has received widespread pushback for telling a female reporter "quiet, piggy" in a video that went viral Monday. He's not going to like this one. The Washington Commanders took on the Miami Dolphins in an NFL overseas game in Spain on Sunday morning. The Dolphins topped the Commanders, 16-13, in overtime. Miami improved to 4-7 on the season with the win, while Washington dropped to 3-8 on the year with the loss. It's been a highly ... Alice and Ellen Kessler ended their lives through medical aid in dying, and "left together, just as they lived: inseparable" "they truly don’t care unless it’s happening to them personally." The move raises questions about how far the administration will go to reshape the armed forces. "Society is not kind to women who age, but aging has given me something that I didn’t even know I needed." Melissa McCarthy turned heads at the Elle Women in Hollywood 2025 event, stepping out at the Four Seasons Hotel looking slimmer than ever. ennifer Aniston and boyfriend Jim Curtis made their red-carpet debu...
(全文较长，仅显示前2000字符，查看完整内容)

Your browser appears to have Javascript disabled. For instructions on how to enable Javascript please click here . If you have any issues, please contact us at challengehelp@humansecurity.com Before we continue... Press & Hold to confirm you are a human (and not a bot). Reference ID d4779312-c4da-11f0-990c-59fc8beac60c

As Large Language Models (LLMs) grow in size and complexity, finding ways to reduce their computational and energy costs has become a critical challenge. One popular solution is quantization, where the precision of parameters is reduced from the standard 16-bit floating-point (FP16) or 32-bit floating-point (FP32) to lower-bit formats like 8-bit or 4-bit. While this approach significantly cuts down on memory usage and speeds up computation, it often comes at the expense of accuracy. Reducing the precision too much can cause models to lose crucial information, resulting in degraded performance. BitNet is a special transformers architecture that represents each parameter with only three values: (-1, 0, 1) , offering a extreme quantization of just 1.58 ( ) bits per parameter. However, it requires to train a model from scratch. While the results are impressive, not everybody has the budget to pre-train an LLM. To overcome this limitation, we explored a few tricks that allow fine-tuning an existing model to 1.58 bits! Keep reading to find out how ! - TL;DR - What is BitNet in More Depth? - Pre-training Results in 1.58b - Fine-tuning in 1.58b - Kernels used & Benchmarks - Conclusion - Acknowledgements - Additional Resources BitNet is an architecture introduced by Microsoft Research that uses extreme quantization, representing each parameter with only three values: -1, 0, and 1. This results in a model that uses just 1.58 bits per parameter, significantly reducing computational and memory requirements. This architecture uses INT8 addition calculations when performing matrix multiplication, in contrast to LLaMA LLM's FP16 addition and multiplication operations. This results in a theoretically reduced energy consumption, with BitNet b1.58 saving 71.4 times the arithmetic operations energy for matrix multiplication compared to the Llama baseline. We have successfully fine-tuned a Llama3 8B model using the BitNet architecture, achieving strong performance on downstream tasks. ...
(全文较长，仅显示前2000字符，查看完整内容)