No it is not possible to train in on the data given through RAG. It will only be able to reference that information. It will pretty much just copy paste relevant information from the documents given.
Thanks for this video! Good job. Perhaps those of you who are trying to use CUDA for GPU use this command CMAKE_ARGS="-DGGML_CUDA=on" FORCE_CMAKE=1 poetry run pip install --no-cache-dir llama-cpp-python as the LLAMA_CUBLAS variable is deprecated and should be used instead GGML_CUDA.
the video that saved my life lmao, ive been wanting to play ark again for a long time but just havent gotting back into it because of all of the "new" stuff, like tek and crap i dont even like that stuff. i miss OG OG ark where it was just building and farming. more PVE like is the way i like ark. shit ass game for pvp anyway. go play rust lmao for that
i still remember the spoiling timer bug in this update wherein you could switch stuff with spoiling timers between your dinos and yourself to stack up the timer :D
Congratulations with "On A Rail"! :) Afaik, for "Overkill" you need to have diamond sword with Smite V (5) and strength potion II (2) and just hit zombie or skeleton with critical hit
I dont think I've ever seen the rail achievement ever done. For reference, you need approximately 375 iron for the rail. Without an iron farm thats a lot!
It's nice view from up here! Maybe you should silently migrate to 1.1 so sheeps can grow back? But if you do, please read about additions and changes in 1.1, maybe you wouldn't want to migrate. Liked the video :)