Want to play around with the Gemma AI open source model launched by Google? This article will teach you how to install it on Windows and Mac computers. The method is very simple. You only need to download the free software LM Studio to do it. You can even play other LLM large language models, and it can run offline without the need for the Internet.
Teach you to install the Google Gemma AI model on your computer and run it without an Internet connection (Windows/Mac)
Click the link above to enter LM Studio. After closing, download your operating system version. It supports Windows, Mac and Linux. Please note that the Mac part is only for Apple Silicon processors:
I use the Windows version as an example. After downloading, you can open it with two clicks and use it without installation:
Although it currently only has an English interface, it is very simple to use. You will see the download point for the Google Gemma AI model (2B) in the middle of the homepage. Google Gemma AI is an open source model, and there are two types: 2B and 7B. The parameters of 2B are 2 billion, and 7B is 7 billion. The 2B file is smaller and does not require computer hardware specifications, so even an older computer can run it easily. Therefore, it is recommended to test 2B first, and then download if you feel it is not enough. 7B try:
Click Download and the download will start. The model file is 2.67GB, and Gemma 7B is almost 4GB:
Click the bottom to view the download progress, the speed depends on your Internet speed:
After downloading, please click on the third Chat icon on the far left:
Then click Select a model to load above. All the AI models you have downloaded will be listed here. I only downloaded Gemma 2B, so there is only this one. Those with enough computer capacity can also download multiple models, so you can switch and test which one is best:
After selecting, this message will pop up. It depends on which one you want to choose:
Then you can start chatting with Gemma AI:
Supports Chinese, but I suggest that if possible, it is better to use English, because this model is very suitable for English tasks, such as creating text, summarization, reasoning, etc. After obtaining the results, they can be translated into Chinese through ChatGPT and the like:
Introduction to Gemma model
The Gemma series is a lightweight, cutting-edge open source model developed based on the same research and technology as the Gemini model. Developed by Google DeepMind and other teams at Google, the name “Gemma” comes from the Latin word meaning “precious stone.” Data support for the Gemma model is provided by development tools that promote innovation, collaboration and the responsible use of artificial intelligence (AI).
Gemma models can run on applications, hardware, mobile devices or cloud services. These models can also be customized by adjusting techniques to make them better at performing important tasks.
Users can not only leverage the Gemma model for text generation, but also adjust the model to focus on the execution of specific tasks. The adjusted Gemma model can bring users more accurate and efficient generative AI solutions.
Source: KOCPC Chinese