Jeremy Howard talks about llama.cpp

[monatis]

I just wanted to bring this to the attention of the community. In the one-and-a-half -hour long lecture, Jeremy Howard, creator of ULMFiT and one of the most reputable figures of language modeling, talks about llama.cpp and demonstrate its usage: A hackers' guide to language models.

Apart from this demonstration, I think that it might be very helpful for those who have a background in software engineering but not in ML, so worth sharing.