vllm.v1.worker.cpu
¶
Modules:
-
attn_utils– -
block_table– -
ec_connector– -
encoder_cache– -
encoder_runner– -
input_batch–CPU implementations for Model Runner V2 input-batch kernels.
-
kv_connector– -
logprob– -
mm_pruning– -
model_states– -
pp_utils–Pipeline-parallel sampled-token synchronization for the CPU runner.
-
rope– -
sampler– -
standalone_model_runner– -
states– -
utils–