@@ -70,6 +70,8 @@ def __init__(
7070 pooling_params : Optional [PoolingParams ] = None ,
7171 preprocess_start_time : Optional [float ] = None ,
7272 preprocess_end_time : Optional [float ] = None ,
73+ inference_start_time : float = 0 ,
74+ llm_engine_recv_req_timestamp : float = 0 ,
7375 multimodal_inputs : Optional [dict ] = None ,
7476 multimodal_data : Optional [dict ] = None ,
7577 disable_chat_template : bool = False ,
@@ -95,8 +97,6 @@ def __init__(
9597 prefill_start_index : int = 0 ,
9698 prefill_end_index : int = 0 ,
9799 num_computed_tokens : int = 0 ,
98- inference_start_time : float = 0 ,
99- llm_engine_recv_req_timestamp : float = 0 ,
100100 ) -> None :
101101 self .request_id = request_id
102102 self .prompt = prompt
@@ -115,6 +115,10 @@ def __init__(
115115 self .arrival_time = arrival_time
116116 self .preprocess_start_time = preprocess_start_time
117117 self .preprocess_end_time = preprocess_end_time
118+ self .inference_start_time = inference_start_time
119+ self .llm_engine_recv_req_timestamp = (
120+ llm_engine_recv_req_timestamp if llm_engine_recv_req_timestamp else time .time ()
121+ )
118122 self .disable_chat_template = disable_chat_template
119123 self .disaggregate_info = disaggregate_info
120124
@@ -163,7 +167,6 @@ def __init__(
163167 self .extend_block_tables = []
164168 # dp
165169 self .dp_rank = dp_rank
166- self .llm_engine_recv_req_timestamp = time .time ()
167170
168171 @classmethod
169172 def from_dict (cls , d : dict ):
0 commit comments