Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
Can't submit QLoRA for eval because `Base model "meta-llama/Llama-2-7b-hf" was not found on hub!`
#241
by
xzuyn
- opened
I could use TheBloke's re-upload, but the file hashes don't seem to match and I'm worried it may not merge correctly. Is it because of the gated access? Could we supply an API key to allow it to work then?
Hi!
We can't use Llama2 models as base models as they have (as you suspected) a gated access mechanism. However, you can merge the weights locally and resubmit as "original" weights.
clefourrier
changed discussion status to
closed