bitcoin
bitcoin

$99134.10 USD 

0.76%

ethereum
ethereum

$3278.52 USD 

-2.15%

tether
tether

$1.00 USD 

0.04%

solana
solana

$253.09 USD 

-0.78%

bnb
bnb

$621.14 USD 

-0.12%

xrp
xrp

$1.42 USD 

21.26%

dogecoin
dogecoin

$0.398650 USD 

3.54%

usd-coin
usd-coin

$1.00 USD 

0.03%

cardano
cardano

$0.966724 USD 

20.51%

tron
tron

$0.199833 USD 

0.29%

avalanche
avalanche

$39.22 USD 

9.50%

shiba-inu
shiba-inu

$0.000025 USD 

0.34%

toncoin
toncoin

$5.46 USD 

-0.74%

sui
sui

$3.42 USD 

-4.33%

bitcoin-cash
bitcoin-cash

$486.65 USD 

0.25%

加密貨幣新聞文章

Google 在 2024 年 Google I/O 大會上推出 Gemini 1.5 Pro、Project Astra、Trillium 等

2024/05/15 18:58

Google I/O 2024 見證了許多值得注意的公告,包括發布具有 100 萬個代幣上下文視窗的 Gemini 1.5 Pro 以及支援 35 種語言的 Gemini Advanced。谷歌也推出了第六代張量處理單元 Trillium,以及能夠記憶和處理視覺資訊的人工智慧助理 Project Astra。此外,Google還展示了 Gemma,這是一個開放語言模型系列,用於創建印度語言的語言模型。

Google 在 2024 年 Google I/O 大會上推出 Gemini 1.5 Pro、Project Astra、Trillium 等

Google Unveils Gemini 1.5 Pro, Project Astra, Trillium, and More at Google I/O 2024

Google 在 2024 年 Google I/O 大會上推出 Gemini 1.5 Pro、Project Astra、Trillium 等

Alphabet Inc.'s Google unveiled a plethora of advancements in artificial intelligence (AI), including Gemini 1.5 Pro, Project Astra, Trillium, and other initiatives, at its annual developer conference, Google I/O, held in Mountain View, California.

Alphabet Inc. 旗下的Google在加州山景城舉行的年度開發者大會 Google I/O 上公佈了人工智慧 (AI) 領域的多項進展,包括 Gemini 1.5 Pro、Project Astra、Trillium 和其他舉措。

Gemini 1.5 Pro

雙子座1.5專業版

Gemini 1.5 Pro, Google's multimodal foundation model, now boasts an expanded context window of two million tokens, allowing it to process and analyze vast amounts of information with unprecedented depth. This advancement enables developers to tackle complex tasks involving large datasets, such as debugging code, extracting insights from unstructured data, and developing next-generation AI applications.

Gemini 1.5 Pro 是 Google 的多模式基礎模型,現在擁有一個包含 200 萬個令牌的擴展上下文窗口,使其能夠以前所未有的深度處理和分析大量資訊。這項進步使開發人員能夠處理涉及大型資料集的複雜任務,例如調試程式碼、從非結構化資料中提取見解以及開發下一代人工智慧應用程式。

Project Astra

阿斯特拉計劃

Project Astra is Google's latest AI assistant that combines computer vision, natural language processing, and memory capabilities. It is designed to understand contextual details, reason about its environment in real time, and interact with users through conversational interfaces. In a demonstration, Project Astra accurately identified objects, recalled previous interactions, and provided relevant information based on visual and audio input from a Google Pixel phone.

Project Astra 是Google最新的人工智慧助手,結合了電腦視覺、自然語言處理和記憶功能。它旨在理解上下文細節,即時推理其環境,並透過對話介面與使用者互動。在演示中,Project Astra 準確地識別了物體,回憶了先前的交互,並根據 Google Pixel 手機的視覺和音訊輸入提供了相關資訊。

Trillium

延齡草

Trillium is Google's sixth-generation tensor processing unit (TPU), a specialized computing architecture optimized for AI training. It delivers a 4.7x improvement in compute performance per chip over the previous generation, enabling more efficient and faster training of large language models. Trillium will be available to Google Cloud customers later this year.

Trillium 是 Google 的第六代張量處理單元 (TPU),這是針對 AI 訓練進行最佳化的專用運算架構。與上一代相比,它的每晶片運算效能提高了 4.7 倍,從而能夠更有效率、更快速地訓練大型語言模型。 Trillium 將於今年稍後向 Google Cloud 客戶推出。

Gemma and Indic Language Initiative

傑瑪和印度語言倡議

Gemma is an open-source family of language models developed by Google. Its technology has been instrumental in empowering Indian developers to create large language models for Indic languages. One such model, Navarasa 2.0, developed by Telugu LLM Labs, is capable of processing content in 15 Indian languages.

Gemma 是由 Google 開發的開源語言模型系列。其技術有助於印度開發人員為印度語言創建大型語言模型。其中一個模型是 Telugu LLM Labs 開發的 Navarasa 2.0,能夠處理 15 種印度語言的內容。

Other Announcements

其他公告

In addition to the aforementioned advancements, Google also announced:

除了上述進步之外,Google還宣布:

  • Axion Processors: Google's custom-designed processors optimized for machine learning tasks.
  • Nvidia's Blackwell GPUs: Nvidia's latest graphics processing units, tailored for AI workloads, available on Google Cloud early in 2025.
  • AI Hypercomputer: A supercomputing architecture designed to maximize efficiency and performance for complex AI tasks.

Google's I/O 2024 showcased the company's commitment to innovation in the field of AI, providing developers with powerful tools and capabilities to push the boundaries of AI applications and enhance user experiences.

Axion 處理器:Google 為機器學習任務而優化的客製化設計處理器。電腦:一種超級運算架構,旨在最大限度地提高效率和效能Google的I/O 2024 展示了該公司對AI 領域創新的承諾,為開發者提供強大的工具和能力,以突破AI 應用的界限並增強使用者體驗。

免責聲明:info@kdj.com

所提供的資訊並非交易建議。 kDJ.com對任何基於本文提供的資訊進行的投資不承擔任何責任。加密貨幣波動性較大,建議您充分研究後謹慎投資!

如果您認為本網站使用的內容侵犯了您的版權,請立即聯絡我們(info@kdj.com),我們將及時刪除。

2024年11月23日 其他文章發表於