Pure CPU run
#9
by
josiah31
- opened
Hello I am new to the AI space (other than using online models and ollama) and I was wondering if anyone would know how to run the model cpu only?
Or is this is not an approachable path to run it?
We have a server that has plenty of cores but we cant afford GPUs at this time.
Our PyTorch code should in principle also support CPU only runs (but it will be slow..). You could also try to incorporate the few lines that changed between this and Llama into ollama etc.