Module http 

Functions

long

Process-wide client with a 120s timeout for model inference calls.

shared

Process-wide client with no explicit timeout.

standard

Process-wide client with a 30s timeout for ordinary API calls.