diff --git a/config.pbtxt b/config.pbtxt index 356f0c0..650c3af 100644 --- a/config.pbtxt +++ b/config.pbtxt @@ -1,7 +1,7 @@ # Triton backend to use name: "base-gemma-3-1b-it" backend: "python" -max_batch_size: 0 +max_batch_size: 4 # Triton should expect as input a single string # input of variable length named 'text_input' @@ -126,7 +126,7 @@ parameters: [ instance_group [ { kind: KIND_AUTO - count: 1 + count: 2 } ]