Michelle ThompsonAccording to the CEO of Chinese tech giant Baidu, more than 70 AI models with more than 1 billion parameters have been issued in China since the country’s new AI legislation went into effect.
According to a Reuters story, Chinese tech company Baidu CEO Robin Li stated on September 5 that more than 70 artificial intelligence (AI) models have been made available in the nation.
The roughly 70 AI language models that have been made public, according to Li, were constructed with over 1 billion parameters, making them technically large language models (LLMs).
Some of the largest LLMs now available on the market include OpenAI’s Chat GPT-4 and Meta’s Llama, which have not officially disclosed their parameter counts but are rumored to have around 170 trillion and 65 billion parameters, respectively.
When it comes to LLMs, parameters are important since more parameters let the model to capture the data’s finer nuances. In the end, this enhances the performance of the models.
On August 30, the public was given access to the most recent version of Baidu’s Ernie chatbot, a well-known AI developer and provider of internet services in China. Following the adoption of China’s new AI rules, it was one of the first companies to make chatbots available for general usage.
Li claims that Ernie 3.5, the most recent version of Baidu’s chatbot, has a processing speed that is twice as fast as the previous edition and a 50% increase in efficiency. He said that the firm intends to release an updated version in the “near future.”
The Chinese tech and e-commerce giant Alibaba debuted two new open-source AI models earlier this summer on August 3, which are supposed to compete with Meta’s Llama 2.
The laws governing AI in China were just recently passed and put into effect. Prior to the new regulations, businesses could run small-scale tests of their AI products, which increased the test size and enabled more features.
