我们来了

$5,000 grew into a forest: American donor returns to China to reunite with desert reclamation partner_我的网站

一 |     

    A group photo from the 1999 donation ceremony, where Ronald Sakolsky presents $5,000 for afforestation efforts. Photo: Courtesy of Luoyang No.2 Foreign Language School
At 2:48 am on early Sunday morning, after a flight of more than 10 hours, Ronald Sakolsky arrived at Beijing Capital International Airport. As he stepped into the arrivals hall, he greeted the waiting reporters with earnest Putonghua "Wo hui lai le," meaning "I'm back."
The thrill of his homecoming had washed over him most vividly during the final moments of his flight. "When I saw the lights in Beijing as we were landing is when I realized that I am back in my second home."
Sakolsky, an American retired teacher from Pennsylvania, is visiting China from Sunday to September 1 for a journey that will take him to Beijing, North China's Inner Mongolia Autonomous Region and Central China's Henan Province. It is a reunion nearly three decades in the making.
That reunion began with a single act of generosity. In 1999, when Sakolsky arrived in China as a participant in a bilateral teacher exchange program, he was moved by Yin Yuzhen and her husband's tireless efforts to curb desertification in the Mu Us Desert in Inner Mongolia. He raised $5,000 to fund their local tree-planting initiative. The young saplings bought with the donation have since matured into more than 50,000 full-grown trees, transforming the once desolate, sandy wasteland into a thriving green landscape.
After decades of lost contact, Yin, a national model worker renowned for her desert restoration achievements, launched an online search for her American benefactor in May, hoping to reconnect with the man who helped launch her lifelong desert reclamation mission.
What followed was a remarkable wave of online support: within less than 48 hours, a nationwide search had tracked Sakolsky down. That swift effort led to their long-awaited virtual reunion, a video call that brought them face to face for the first time in 27 years.
Asked what he is most looking forward to about this journey, Sakolsky told the Global Times that "I want to see Ms Yin, see the forests, see the students I have not seen for almost 30 years, my colleagues. I want to see the Chinese people." 
The trip itself felt like a dream he never thought would come true, he added. 
Now, the dream is about to come true. Later on Sunday, Sakolsky will fly to Ordos, Inner Mongolia, the site of the transformative environmental project that bound his fate with Yin's all those years ago.
During his China visit, Sakolsky will join Yin in a new tree-planting activity. He will also hold exchanges and discussions with teachers and students from the Luoyang No.2 Foreign Language School where he once taught, and participate in official China-US teacher exchange events, according to the Xinhua News Agency. 
As he prepares to see the transformed landscape with his own eyes, Sakolsky shared with the Global Times about how much China itself has changed over the past two decades. From photographs, he said, he had already seen the barren desert he remembered turn into lush green forests; he had also heard that Luoyang has become far more vibrant and beautiful.  
"Hopefully the message gets out to the world that Chinese and Americans can work together and make a peaceful world," he said.
。    8月26日消息，SuperCLUE Image放出8月份中文文生图测评榜单。    这次一共测了14款国内外主流模型，还新增复杂信息布局这个考核维度，专门看模型做信息图、分镜、UI界面的能力。    这次榜单最大看点，商汤SenseNova U1 Pro拿到93.81分，以0.58分的微弱优势超过OpenAI GPT Image 2，直接拿下全球总榜第一。     字节跳动Doubao Seedream 5.0 pro、阿里Qwen Image 3.0 Pro紧随其后。

二 | 总榜前五里面，国内模型占了3个。

三 |          从得分情况能看出来，现在不少模型做画面还原、创意生成已经很成熟，这两项平均分接近90分。    真正拉开差距的是图文一致性，还有刚新增的复杂信息布局，两项平均分只有72分出头。特别是复杂信息布局，最高分和最低分差74.2分，是分化最厉害的一项。

四 |     简单说，生成好看图片的大模型很多，但能在一张图里搞定大量文字、表格、流程图的模型很少。    榜单也能看出明显两极分化。排名前2的商汤、OpenAI模型，六大测评维度全部站在第一梯队，属于全能选手。

五 |     但很多中尾部模型偏科很严重，某一项能力很强，别的维度直接拉胯，只靠单个亮点已经冲不上高排名。    国产模型不是所有项目都占优。汉字生成、现实复现这两块国内表现更好，可在图文一致性上，和海外模型还有16.56分的不小差距。    总的来说，国产头部文生图已经站上全球第一的位置，但不是整体全部领先。能不能处理复杂图文信息，已经变成判断文生图模型好不好用的核心标准。                   

。

Current article:http://bfdjs1.suilaoheirenzhaizhaikang.sbs/ea0o/3b22e.html

Published on:09:17:54