Open jkfnc opened 1 month ago
is this a question, a statement or what... ?
Sorry, its a question, since LLama 3.1 needs RoPE scaling is there any changes needed to Cake for it to work .
RoPE is implemented here https://github.com/evilsocket/cake/blob/8a782b8a76f5ae1c3438c0853c9ecd34306d4197/cake-core/src/models/llama3/attention.rs#L25 ... to be honest i didn't have time yet to check the structural differences with LLama 3.1
Llama 3.1 got released, anything needs to change to get it working.