[en] Add protocols

docs/zh/docs/en/blogs/0327-transformer.md

@@ -20,7 +20,7 @@ While the Transformer has thrived, some dissenting voices have emerged, such as:
 
 This is because the power of the Transformer is also its weakness: the inherent self-attention mechanism in the Transformer presents challenges, primarily due to its quadratic complexity. This complexity makes the architecture **computationally expensive and memory-intensive** when dealing with long input sequences or in resource-constrained situations.
 
-In simple terms, this means that as the sequence length (for example, the number of words in a paragraph or the size of an image) processed by the Transformer increases, the required computational power grows quadratically, quickly becoming enormous. Hence, there is a saying that "the Transformer is not efficient." This is also a major reason for the global shortage of computing power triggered by the current AI boom.
+In simple terms, this means that as the sequence length (for example, the number of words in a paragraph or the size of an image) processed by the Transformer increases, the required computational power grows quadratically, quickly becoming enormous. Hence, there is a saying that "the Transformer is not efficient." This is also a major reason for the global shortage of computing resource triggered by the current AI boom.
 
 Based on the limitations of the Transformer, many non-Transformer architectures have emerged, including China's RWKV, Meta's Mega, Microsoft's RetNet, Mamba, and DeepMind's Hawk and Griffin. These models have been proposed following the dominance of the Transformer in the LLM development landscape.
 
@@ -94,7 +94,7 @@ RWKV, the most representative non-Transformer research, has made significant pro
 
 However, several investors have told AI Technology Review that they have struggled with whether to invest in RWKV, betting on non-Transformers. Due to significant internal disagreements—fearing that non-Transformers may not perform well—they ultimately gave up.
 
-Currently, based on the existing hardware computing power foundation, it is very challenging to create LLMs on the edge with Transformers; calculations and inferences still need to be done in the cloud, and the response speed is unsatisfactory, making it difficult for end-users to accept.
+Currently, based on the existing hardware computing resource foundation, it is very challenging to create LLMs on the edge with Transformers; calculations and inferences still need to be done in the cloud, and the response speed is unsatisfactory, making it difficult for end-users to accept.
 
 An industry insider told AI Technology Review, "On the edge, RWKV may not necessarily be the optimal solution, because with advancements in semiconductors, AI chips are evolving. In the future, the costs of hardware, computing, and energy will eventually be leveled out, and LLMs could easily run directly on the edge without needing significant changes to the underlying architecture. One day, we will reach such a critical point."
 

docs/zh/docs/en/blogs/0403-cp-to-profit.md

@@ -3,36 +3,36 @@ hide:
   - toc
 ---
 
-# "AI Process Orchestration" Transforms Computing Power into "Computational Benefits"
+# "AI Process Orchestration" Transforms computing resource into "Computational Benefits"
 
 !!! info
 
-    National-level specialized and innovative "Little Giant" enterprise launches computing power ecological platform
+    National-level specialized and innovative "Little Giant" enterprise launches computing resource ecological platform
 
     Source: [Jiefang Daily](https://www.shobserver.com/staticsg/res/html/journal/detail.html?date=2024-04-02&id=370048&page=07); Reporter: Yu Taoran
 
-The rise of large AI models has made GPU (Graphics Processing Unit) computing power an extremely important resource. In this field, GPU computing power cloud services play a crucial role, enabling the free flow of computing resources across regions and industries. Recently, the national-level specialized and innovative "Little Giant" enterprise, Shanghai Daoke, in collaboration with industry partners, launched the "d.run Integrated Computing Power Solution." This solution includes computing center services in Shanghai, Hefei, and other locations, algorithm development, model fine-tuning tools, and application development tools such as intelligent Q&A and process orchestration, providing users with an AI computing power ecological platform.
+The rise of large AI models has made GPU (Graphics Processing Unit) computing resource an extremely important resource. In this field, GPU computing resource cloud services play a crucial role, enabling the free flow of computing resources across regions and industries. Recently, the national-level specialized and innovative "Little Giant" enterprise, Shanghai DaoCloud, in collaboration with industry partners, launched the "d.run Integrated Computing Solution." This solution includes computing center services in Shanghai, Hefei, and other locations, algorithm development, model fine-tuning tools, and application development tools such as intelligent Q&A and process orchestration, providing users with an AI computing resource ecological platform.
 
 [![Jiefang Daily Front Page](./images/profit01.png)](https://www.shobserver.com/staticsg/res/html/journal/detail.html?date=2024-04-02&id=370048&page=07)
 
 ## AI Infrastructure
 
-In the past five years, China's computing power industry has experienced rapid growth, with an average annual growth rate exceeding 30%. However, challenges still exist, such as a lack of core chips and a weak domestic basic software ecosystem, insufficient transmission capacity, and an increase in ineffective computing power. Chen Qiyan, founder and CEO of Shanghai Daoke, believes that after the rise of LLMs, in addition to the insufficient supply of GPU chips, China's AI industry also faces issues in collaborative services and practical applications. How to improve the efficiency of existing domestic computing power? This requires the joint efforts of all enterprises in the computing power industry chain, including cloud service providers, to create a comprehensive integrated computing power solution.
+In the past five years, China's computing industry has experienced rapid growth, with an average annual growth rate exceeding 30%. However, challenges still exist, such as a lack of core chips and a weak domestic basic software ecosystem, insufficient transmission capacity, and an increase in ineffective computing resource. Chen Qiyan, founder and CEO of Shanghai DaoCloud, believes that after the rise of LLMs, in addition to the insufficient supply of GPU chips, China's AI industry also faces issues in collaborative services and practical applications. How to improve the efficiency of existing domestic computing resource? This requires the joint efforts of all enterprises in the computing industry chain, including cloud service providers, to create a comprehensive Integrated Computing Solution.
 
-In 2014, Chen Qiyan's team from the EMC China Research Institute of Yihuanxin embarked on an exploration of AI infrastructure, focusing on the utilization of computing power. Their goal was to develop a platform capable of orchestrating and optimizing computing resources effectively. This journey led to the establishment of Shanghai Daoke, which has since evolved into a national-level "Little Giant" enterprise specializing in computing power services.
+In 2014, Chen Qiyan's team from the EMC China Research Institute of Yihuanxin embarked on an exploration of AI infrastructure, focusing on the utilization of computing resource. Their goal was to develop a platform capable of orchestrating and optimizing computing resources effectively. This journey led to the establishment of Shanghai DaoCloud, which has since evolved into a national-level "Little Giant" enterprise specializing in computing services.
 
 ## The d.run Platform
 
-The newly launched d.run Integrated Computing Power Solution aims to provide a comprehensive suite of services that cater to the growing demands of AI applications. The platform integrates various resources, including computing centers located in strategic areas like Shanghai and Hefei, to ensure high availability and scalability. It also offers tools for algorithm development and model fine-tuning, making it easier for businesses to harness AI technology without needing extensive in-house expertise.
+The newly launched d.run Integrated Computing Solution aims to provide a comprehensive suite of services that cater to the growing demands of AI applications. The platform integrates various resources, including computing centers located in strategic areas like Shanghai and Hefei, to ensure high availability and scalability. It also offers tools for algorithm development and model fine-tuning, making it easier for businesses to harness AI technology without needing extensive in-house expertise.
 
-The platform's intelligent Q&A system and process orchestration capabilities allow users to streamline operations and enhance productivity. By automating processes and facilitating better resource management, the d.run platform transforms raw computing power into tangible computational benefits, enabling organizations to focus on innovation rather than infrastructure.
+The platform's intelligent Q&A system and process orchestration capabilities allow users to streamline operations and enhance productivity. By automating processes and facilitating better resource management, the d.run platform transforms raw computing resource into tangible computational benefits, enabling organizations to focus on innovation rather than infrastructure.
 
 ## Industry Collaboration
 
-Chen Qiyan emphasizes the importance of collaboration within the industry to overcome existing challenges. By working together with other enterprises, cloud service providers, and technology partners, Shanghai Daoke aims to create a robust ecosystem that drives the advancement of AI technologies in China. This collaborative approach is essential for addressing the issues related to GPU supply, enhancing software ecosystems, and maximizing the efficient use of computing resources.
+Chen Qiyan emphasizes the importance of collaboration within the industry to overcome existing challenges. By working together with other enterprises, cloud service providers, and technology partners, Shanghai DaoCloud aims to create a robust ecosystem that drives the advancement of AI technologies in China. This collaborative approach is essential for addressing the issues related to GPU supply, enhancing software ecosystems, and maximizing the efficient use of computing resources.
 
-As the demand for AI solutions continues to grow, the d.run platform positions itself as a vital player in the computing power landscape, providing businesses with the tools and resources necessary to leverage AI effectively. The initiative is expected to not only boost the efficiency of computing power utilization but also contribute to the overall growth of the AI industry in China.
+As the demand for AI solutions continues to grow, the d.run platform positions itself as a vital player in the computing resource landscape, providing businesses with the tools and resources necessary to leverage AI effectively. The initiative is expected to not only boost the efficiency of computing resource utilization but also contribute to the overall growth of the AI industry in China.
 
 ## Conclusion
 
-The launch of the d.run Integrated Computing Power Solution marks a significant step forward in the evolution of AI infrastructure in China. By transforming computing power into computational benefits, Shanghai Daoke and its partners are paving the way for a more efficient and innovative AI ecosystem. As the industry continues to evolve, the emphasis on collaboration and integration will be crucial in addressing the challenges and unlocking the full potential of AI technologies.
+The launch of the d.run Integrated Computing Solution marks a significant step forward in the evolution of AI infrastructure in China. By transforming computing resource into computational benefits, Shanghai DaoCloud and its partners are paving the way for a more efficient and innovative AI ecosystem. As the industry continues to evolve, the emphasis on collaboration and integration will be crucial in addressing the challenges and unlocking the full potential of AI technologies.

docs/zh/docs/en/blogs/0408-after-kimi.md

@@ -30,7 +30,7 @@ Wenxin Yiyan also announced an upgrade scheduled for April, which will include l
 
 However, many industry insiders are skeptical about the major companies' attempts to "follow suit." They believe that Kimi's leading position in the long text domain will not be easily surpassed.
 
-Perhaps for this reason, in February of this year, during Moonlight Dark Side's latest round of financing, Alibaba, as the lead investor, converted 70-80% of its investment into computing power services.
+Perhaps for this reason, in February of this year, during Moonlight Dark Side's latest round of financing, Alibaba, as the lead investor, converted 70-80% of its investment into computing services.
 
 ## The Soul of Kimi
 
@@ -44,7 +44,7 @@ During his doctoral studies, Yang Zhilin published two significant works, Transf
 
 Transformer-XL became the first attention-based language model to fully surpass RNNs, while XLNet received best paper nominations at 20 top conferences.
 
-Yang Zhilin and his team created Kimi's outstanding lossless compression technology. This technology reduces the storage needs of parameters, inference computing power, and bandwidth usage for data transmission, enabling efficient lossless processing of millions of long tokens.
+Yang Zhilin and his team created Kimi's outstanding lossless compression technology. This technology reduces the storage needs of parameters, inference computing resource, and bandwidth usage for data transmission, enabling efficient lossless processing of millions of long tokens.
 
 In contrast, to quickly catch up with the long text trend, other major companies have had to settle for using retrieval-augmented generation (RAG) technology.
 
@@ -92,11 +92,11 @@ For example, for product growth positions, candidates are required to have exper
 
 "When you can't beat them, join them." Despite having its own Tongyi Qianwen, Alibaba has also provided support for Kimi.
 
-Currently, Kimi has borrowed Alibaba's Nvidia (NVDA.O) A800 and A100 GPU processors for expansion, and future support for Kimi's computing power will mainly come from Alibaba.
+Currently, Kimi has borrowed Alibaba's Nvidia (NVDA.O) A800 and A100 GPU processors for expansion, and future support for Kimi's computing resource will mainly come from Alibaba.
 
-In February of this year, during Moonlight Dark Side's latest round of investment exceeding $1 billion, Alibaba, as the lead investor, converted 70-80% of its investment into computing power servers.
+In February of this year, during Moonlight Dark Side's latest round of investment exceeding $1 billion, Alibaba, as the lead investor, converted 70-80% of its investment into computing resource servers.
 
-With Alibaba's support, Kimi no longer needs to worry about downtime due to insufficient computing power. An insider has also indicated that the expansion will not be done all at once. Rapid expansion can lead to idle and wasted computing power, so a certain strategy is needed. For example, Kimi will also predict user usage patterns.
+With Alibaba's support, Kimi no longer needs to worry about downtime due to insufficient computing resource. An insider has also indicated that the expansion will not be done all at once. Rapid expansion can lead to idle and wasted computing resource, so a certain strategy is needed. For example, Kimi will also predict user usage patterns.
 
 Since the release of ChatGPT in November 2022, there have been over 200 AI models in China, and the number continues to increase. The emergence of Kimi has awakened a sense of crisis among industry giants.
 

docs/zh/docs/en/blogs/0429-ai-survey.md

@@ -6,7 +6,7 @@ In March of this year, ClearML released the results of a global AI survey conduc
 
 The survey focuses on gaining more insights into global AI infrastructure plans, including respondents':
 
-1. Computing power infrastructure growth plans;
+1. Computing Infrastructure growth plans;
 2. Current experiences with scheduling and computing solutions;
 3. Usage and plans for models and AI frameworks in 2024.
 

docs/zh/docs/en/blogs/0514-gpt4o.md

@@ -244,7 +244,7 @@ Finally, a huge thank you to the team for their tremendous efforts to achieve th
 
 ![Image](./images/gpt4o-34.webp)
 
-It’s worth mentioning that last week, Altman stated in an interview that while universal basic income may be difficult to achieve, we can realize "universal basic compute." In the future, everyone could access GPT’s computing power for free, which could be used, resold, or donated.
+It’s worth mentioning that last week, Altman stated in an interview that while universal basic income may be difficult to achieve, we can realize "universal basic compute." In the future, everyone could access GPT’s computing resource for free, which could be used, resold, or donated.
 
 "The idea is that as AI becomes more advanced and embedded in every aspect of our lives, owning a unit of a large language model like GPT-7 may be more valuable than money; you own a part of productivity," explained Altman.
 

docs/zh/docs/en/blogs/index.md

@@ -39,18 +39,18 @@ This channel will closely follow technology trends and collect news from the AI
     Introducing DBRX, an open general LLM created by Databricks.
     In a series of standard benchmark tests, DBRX has set new technical standards among established open LLMs.
 
-- [“AI Process Orchestration” Turns Computing Power into “Profit”](./0403-cp-to-profit.md)
+- [“AI Process Orchestration” Turns computing resource into “Profit”](./0403-cp-to-profit.md)
 
-    The national-level specialized and innovative "little giant" enterprise DaoCloud has launched the computing power ecosystem platform d.run.
+    The national-level specialized and innovative "little giant" enterprise DaoCloud has launched the computing resource ecosystem platform d.run.
 
 - [Who Will Replace the Transformer?](./0327-transformer.md)
 
     The Transformer started with Google's 2017 paper "Attention Is All You Need,"
     but why has it been popularized by OpenAI and dominated the field?
     What common challenges do non-Transformer models face?
 
-- [The Financial Industry Enters the Era of LLMs, Computing Power Infrastructure Becomes the Key to Victory](./0326-compute-power.md)
+- [The Financial Industry Enters the Era of LLMs, Computing Infrastructure Becomes the Key to Victory](./0326-compute-power.md)
 
-    DaoCloud is leading a seminar on computing power and LLM industries organized by the local financial industry in Shanghai.
+    DaoCloud is leading a seminar on computing resource and LLM industries organized by the local financial industry in Shanghai.
 
 > Contributions and reprints in various forms are welcome.

docs/zh/docs/en/index.md

@@ -24,7 +24,7 @@ d.run supports account registration via mobile phone and email.
 
 1. After successful registration, you will return to the login page, enter your username or email, and log in to your account.
 
-You can now start purchasing computing power.
+You can now start purchasing computing resource.
 
 ![market](./images/regis01.png)