Meta releases the latest open-source big model Llama 3
日微牧
发表于 2024-4-19 14:40:01
1251
0
0
On April 18th local time, Meta released its latest open-source large model Llama 3, providing pre trained and instruction fine-tuning versions of 8B and 70B.
According to Meta's official introduction, Llama 3 is trained on two customized 24K GPU clusters based on data exceeding 15T, which is 7 times larger than the dataset used by Llama 2 and 4 times more code. Additionally, Llama 3 supports 8K context length, which is twice the capacity of Llama 2. Moreover, Meta has also released parameter comparisons between two versions of Llama 3 and competitors such as Google Gemma, Google Gemini, Mistral, and Antioptic's Claude 3.
Meta CEO Zuckerberg introduced that Meta AI will be integrated into the search boxes at the top of Meta's major products WhatsApp, Instagram, Facebook, and Messenger, and will create a website called meta.ai to make it easier to use.
Meta Chief Scientist Yann LeCun stated that Llama will continue to release more versions in the coming months and will soon release research papers on Llama 3. Jim Fan, a senior scientist at Nvidia, believes that the possible release of Llama 3-400B and above versions in the future will become a "watershed", and the open source community will be able to use GPT-4 level models.
CandyLake.com 系信息发布平台,仅提供信息存储空间服务。
声明:该文观点仅代表作者本人,本文不代表CandyLake.com立场,且不构成建议,请谨慎对待。
声明:该文观点仅代表作者本人,本文不代表CandyLake.com立场,且不构成建议,请谨慎对待。
猜你喜欢
- Robin Lee: The average daily adjustment amount of Wenxin Model exceeded 1.5 billion, 30 times more than that of a year ago
- Will DeepMind's open-source biomolecule prediction model win the Nobel Prize and ignite a wave of AI pharmaceuticals?
- "AI new generation" big model manufacturer Qi "roll" agent, Robin Lee said that it will usher in an era of "making money by thinking"
- Robin Lee said that the illusion of the big model has basically eliminated the actual measurement of ERNIE Bot?
- AI Weekly | Yang Zhilin claims that Kimi has over 36 million monthly active users; Robin Lee: The illusion of big model is basically eliminated
- ERNIE Bot has more than 400 million users, Baidu Wu Tian: the big model is reshaping the industrial intelligence engine
- In October of this year, Tesla Model Y won the sales championship for first tier and new first tier city models
- Alibaba CEO Wu Yongming: AI development requires a batch of open-source models of different scales and fields
- Baidu's Q3 core net profit increased by 17%, exceeding expectations. Wenxin's large model daily usage reached 1.5 billion
-
知名做空机构香橼研究(Citron Research)周四(11月21日)在社交媒体平台X上发布消息称,该公司已决定做空“比特币大户”微策略(Microstrategy)这家公司,并认为该公司已经将自己变身成为一家比特币投资基金 ...
- caffycat
- 12 小时前
- 支持
- 反对
- 回复
- 收藏
-
每经AI快讯,11月20日,文远知行宣布旗下自动驾驶环卫车S6与无人扫路机S1分别在新加坡滨海湾海岸大道与滨海艺术中心正式投入运营。据介绍,这是新加坡首个商业化运营的自动驾驶环卫项目。 ...
- star8699
- 前天 19:48
- 支持
- 反对
- 回复
- 收藏
-
上证报中国证券网讯(记者王子霖)11月20日,斗鱼发布2024年第三季度未经审计的财务报告。本季度斗鱼依托丰富的游戏内容生态,充分发挥主播资源和新业务潜力,持续为用户提供高质量的直播内容及游戏服务,进一步 ...
- goodfriendboy
- 前天 20:09
- 支持
- 反对
- 回复
- 收藏
-
人民网北京11月22日电 (记者栗翘楚、任妍)2024广州车展,在新能源汽车占据“半壁江山”的同时,正加速向智能网联新能源汽车全面过渡,随着“端到端”成为新宠,智能驾驶解决方案成为本届广州车展各大车企竞 ...
- 3233340
- 6 小时前
- 支持
- 反对
- 回复
- 收藏