3 ms·
That hasn’t been true for a while. Falcon 40B seemingly outperforms LLaMA 60B according to the OpenLLM leaderboard https://huggingface.co/tiiuae/falcon-40b htt
by kirill5pol 3y ago
That hasn’t been true for a while. Falcon 40B seemingly outperforms LLaMA 60B according to the OpenLLM leaderboard
https://huggingface.co/tiiuae/falcon-40b https://huggingface.co/tiiuae/falcon-40b
- lolinder 3y agoFair enough. I haven't really looked at Falcon as a replacement for LLaMA yet because it isn't supported by llama.cpp, but it looks promising.