I tried to finetune CodeLLama 7B with my own codebase using the
finetuning pipeline. However, even with a Standard_NC24rs_v3 compute
(NVidia Tesla V100 16 GB vRam) I get an OOM error. Considering that this
compute is already almost $18/hour, does this mean that I need an ever
bigger compute to finet...
This will be available on December 7th. If you are interested in trying
this sooner, please get back to me with your Azure AAD ID (entra id) and
Azure subscription ID in a private message. Thanks Manoj
Latest Comments