Chen Tianqi officially announced the new APP, which allows the mobile phone to run the large model natively, and the app store can be directly downloaded and used

Available on iPhone and iPad.

Some time ago, the author of TVM, MXNET, XGBoost, assistant professor of CMU, OctoML CTO Chen Tianqi and other researchers jointly developed a project that ignited the AI world.

The project, called MLC-LLM, lets you "compile and run large language models on any device," including mobile devices (like the iPhone), consumer computers (like the Mac), and web browsers. The running effect is as follows:

In less than two days since MLC-LLM was launched, the number of stars on GitHub is close to one thousand. On social networks, people praised it one after another, saying that MLC-LLM "let me experience the horror of running AI with low precision." As of now, the project has gained 7.5k stars on Github.

Just now, Chen Tianqi announced another good news: MLC Chat app has been launched in Apple's App Store.

MLC Chat is an app that enables you to chat natively with Open Language Models on iPad and iPhone, and is part of the MLC LLM project.

After MLC Chat downloads the model into the app, everything runs locally, no server support is required, and no internet connection is required, no information is logged.

After hearing the news, we downloaded the app to try it out.

Open the APP, you can see that the app has pre-installed a model - RedPajama.

Opening this model, we will directly enter the chat window, where we can have a direct conversation with RedPajama.

If you are not satisfied with this model, you can add a new model by entering the URL on the home page. However, this process of downloading new models can be a bit slow.

Since the app has just been released, we haven't seen feedback from other users yet.

However, what is certain is that with the emergence and popularity of such applications, the threshold for using large language models will continue to decrease. In the past, only researchers and technologists in specialized fields were able to effectively utilize these models for various tasks. However, now with the development of open source projects and development tools, more and more people can easily use these models.

If you are also interested in MLC Chat, you can go to Apple's APP Store to download and experience it.

Reference link:

MLC Chat download address: https://apps.apple.com/us/app/mlc-chat/id6448482937