1-click to GCP Vertex AI Endpoint fails
#10
by
aflansburg
- opened
1-click deployment to Vertex endpoint does not work.
The error seems fairly straightforward. The model is stored successfully in the registry, but this error is seen when the server shard is coming up for the endpoint. This is from the Shard complete standard error output:
raise LocalEntryNotFoundError(\\n\\n[rank1]: huggingface_hub.utils._errors.LocalEntryNotFoundError: File consolidated.safetensors of model mistralai/Mistral-Nemo-Base-2407 not found in /tmp. Please run `text-generation-server download-weights mistralai/Mistral-Nemo-Base-2407` first.
I'm unfamiliar with how you use safetensors in Vertex. Assuming it's more involved than a 1-click process :)
Hello @aflansburg ,
Can you share your whole error stack?