Large Language Model (LLM) API#
LLM processor builder#
Build a processor using the given config. |
Processor#
A processor is composed of a preprocess stage, followed by one or more processing stages, and finally a postprocess stage. |
Processor configs#
The processor configuration. |
|
The configuration for the HTTP request processor. |
|
The configuration for the vLLM engine processor. |
|
The configuration for the SGLang engine processor. |
Stage configs#
The configuration for the chat template stage. |
|
The configuration for the tokenizer stage. |
|
The configuration for the detokenize stage. |
|
The configuration for the prepare multimodal stage. |