Meta today unveiled Llama 2, its next generation large language model, that is fully open source, free and available for research and commercial use. Here's how you can easily get started with Llama 2 and give Llama-2-chat a try right now.
Hugging Chat
Try out the 70B Chat model for free with super fast inference, web search, and powered by open-source tools!
Llama 2 on Perplexity
Perplexity Labs' Llama Chat implementation is super fast (think GPT 3.5 Turbo), however for now it is limited to the 7B paramater model.
Llama 2 Playground
a16z in collaboration with Replicate have created the Llama 2 playground where you can interact with the finetuned 7B and 13B parameter models in your browser. Simply visit https://www.llama2.ai to get started.
Llama 2 Chatbot App
You can also clone this chatbot app as a starting point and run the app with Streamlit.
Replicate
Replicate allows you to run models in the cloud at scale. You can use the a16z's public repos to access the Replicate API directly, or use the Cog template and push your own model to Replicate.
Meta AI
Finally, you can download the model weights and tokenizer directly from Meta, however you must complete a license form and agree to their privacy policy. Once your request is approved, you will receive a signed URL over email. Then run the download.sh script, passing the URL provided when prompted to start the download.
The launch of Llama 2 opens up thrilling new possibilities in open source large language models. As developers get hands-on with the technology, we can expect rapid iteration and advancement. Stay tuned as the community puts LLaMA 2 through its paces, uncovering its potential, strengths, and areas for improvement.