Released the "Wen Xin Yiyan" artificial intelligence dialogue robot product in the past two months, Chinese technology giants Baidu low -key announcement announced the progress of large -scale enterprise models.
According to Caixin.com, Baidu's company Baidu Cloud on Tuesday (May 9) shows the application of large models on the enterprise.Compared with the high -profile warm -up of the "Wen Xin Yizhi" press conference on March 16, the announcement of corporate products is more low -key. Previously, a closed door for corporate customers was held on March 27.
At this exchange meeting, Baidu Cloud announced the big model platform "Wenxin Qianfan" for corporate customers. The platform provides enterprises with data labeling processing, re -training fine -tuning, large model compression compression, Prompt automation, the development of large modeling functions, as well as large model custody and online reasoning services.The platform also supports external plug -ins, as well as the introduction of large models of third -party models. Baidu Cloud defines the platform as the production and distribution center of large models.
Baidu Cloud demonstrated the data analysis ability of "Wenxin Qianfan" on the spot. By entering text instructions, the platform will demonstrate the data set as a chart, but it fails to achieve the transformation from the cake chart to the folding line chart according to the instruction.Function.In addition, Baidu also demonstrated data import and re -training function. By importing industry data, this model can process specific tasks.
In the deployment model, in addition to providing services such as reasoning, fine -tuning, and hosting on the public cloud, Baidu Cloud also provides a privatization deployment plan for software authorization, software and hardware, and leasing services.
According to reports, the ability of large models has become an important selling point for Chinese Internet cloud manufacturers, and corporate related products have made frequent progress.On April 18, Alibaba's office collaboration software nails announced that they took the lead in accessing Ali's "Qianwen" model and promoted the implementation of the big model in the industry. On the same day, the byte beating also released a large model training cloud platform to provide enterprises with providing enterprises withLarge model computing power services; Tencent also announced on April 14 that it launched the worldwide computing power cluster equipped with Nvidia's latest GPU H800, providing underlying computing services for large models.
