- In
clients.baml I have this configuration:
client<llm> CustomLlama31_405B {
provider "openai-generic"
options {
default_role "user"
model "meta/llama-3.1-405b-instruct"
api_key "" // No API key needed for internal endpoint
base_url "http://localhost:9494/v1"
headers {
"Host" "llama3.dgx-k8s.production"
"Content-Type" "application/json"
}
temperature 0.8
max_tokens 3000
}
- When I click the "Run..." button on the prompt in BAML Playground in VSCode, I receive an error. When I copy the Curl and paste it into terminal, it works fine.
The raw output suggest that the request header, Host key/value isn't being passed when I click the "Run ..." button