Hello.
I would like to start off by thanking you for all the work you are doing for making smaller and faster models, allowing those with lesser hardware to try experimenting with local LLMs. I am working with an older laptop (year 2013) which has no gpu, so I use the cpu backend via llama.cpp. I was wondering if it would be possible to fine tune a smaller model with your antidoom through colab or kaggle using their free gpu resources. And if so, could a guide be made showing how to utilize those services to apply antidoom for those who might be without gpu?
Hello.
I would like to start off by thanking you for all the work you are doing for making smaller and faster models, allowing those with lesser hardware to try experimenting with local LLMs. I am working with an older laptop (year 2013) which has no gpu, so I use the cpu backend via llama.cpp. I was wondering if it would be possible to fine tune a smaller model with your antidoom through colab or kaggle using their free gpu resources. And if so, could a guide be made showing how to utilize those services to apply antidoom for those who might be without gpu?