Just last night, Meta released the latest big open source model that can break bread with OpenAI: Llama 3.1.
There are three versions of the model:
- 8B
- 70B
- 405B
For this release, Meta has evaluated performance on over 150 benchmark datasets covering a wide range of languages. In addition, Meta has conducted extensive manual evaluations comparing Llama 3.1 to competing models in real-world scenarios.Meta's experimental evaluations have shown that:
Evaluation data for the 405B version is decently competitive with the strongest current GPT-4, GPT-4o and Claude 3.5 Sonnet:
In addition, the 8B and 70B versions perform equally well or even better compared to models with similar parameter scales:
Running Llama 3.1 with ollama
Here's a taste of what you can do to get Llama 3.1 version 8B running locally by hand. Just follow the steps below and you don't need to have any technical background to do it.
- Install ollama
Go to the official website to download the latest Ollama:/
-
Once the download is complete, open the installer and complete the installation steps.
-
Open a terminal and execute the command:
ollama run llama3.1
Waiting for the model to finish downloading
- Complete the model download as shown below:
Already popping up tips:Send a message, can now interact with the AI.
- Test it out and try to get him to teach me Java?
colored egg
Finally, the routine, ask who you are:
What's going on? Is it ChatGPT? Do any of you know why? If you know, please ask in the comment section.
Welcome to my public number: program ape DD. the first time to understand the cutting-edge industry news, share in-depth technical dry goods, access to high-quality learning resources