You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Llama 2 or derivative works thereof).
Why doesn’t Meta AI want me to use Llama 2 outputs to train ZachNet?
You are allowed if ZachNet is based on Llama 2, so I’d guess “because non-llama-based ZachNet wouldn’t have the usage restrictions, or the publicity benefits for Meta”.
Why doesn’t Meta AI want me to use Llama 2 outputs to train ZachNet?
You are allowed if ZachNet is based on Llama 2, so I’d guess “because non-llama-based ZachNet wouldn’t have the usage restrictions, or the publicity benefits for Meta”.