Index

Symbols

--block-size command line option --disable-log-stats command line option --download-dir command line option --dtype command line option --enable-prefix-caching command line option --gpu-memory-utilization command line option --load-format command line option --max-model-len command line option --max-num-batched-tokens command line option --max-num-seqs command line option --max-paddings command line option --max-parallel-loading-workers command line option

--model command line option --pipeline-parallel-size command line option --quantization command line option --revision command line option --seed command line option --swap-space command line option --tensor-parallel-size command line option --tokenizer command line option --tokenizer-mode command line option --tokenizer-revision command line option --trust-remote-code command line option --worker-use-ray command line option

A

abort() (vllm.engine.async_llm_engine.AsyncLLMEngine method) abort_request() (vllm.engine.llm_engine.LLMEngine method)

add_request() (vllm.engine.llm_engine.LLMEngine method) AsyncLLMEngine (class in vllm.engine.async_llm_engine)

C

command line option --block-size --disable-log-stats --download-dir --dtype --enable-prefix-caching --gpu-memory-utilization --load-format --max-model-len --max-num-batched-tokens --max-num-seqs --max-paddings --max-parallel-loading-workers --model --pipeline-parallel-size --quantization --revision --seed --swap-space --tensor-parallel-size --tokenizer --tokenizer-mode --tokenizer-revision --trust-remote-code --worker-use-ray

G

generate() (vllm.engine.async_llm_engine.AsyncLLMEngine method)

L

LLMEngine (class in vllm.engine.llm_engine)

M

module vllm.engine vllm.sampling_params.SamplingParams

S

step() (vllm.engine.llm_engine.LLMEngine method)

V

vllm.engine module

vllm.sampling_params.SamplingParams module