5fa1a76
1
2
3
Next, the timeout is reset on every queue fetch, meaning you could wait much more than 1ms before running the inference (delaying the first request by that much).
