apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 00:33:03 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

适配vLLM无arrival_time；适配vLLM model必传；RequestFuncInput/RequestFuncOutput/SampleRequest新增用例编号no

Browse Source
This commit is contained in:
lijingning
2025-07-15 16:53:03 +08:00
committed by Zhang Yulong
parent 1b712bba82
commit 9d6a42b334
3 changed files with 215 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
benchmarks/benchmark_dataset.py
View File

@@ -38,7 +38,7 @@ class SampleRequest:
"""
Represents a single inference request for benchmarking.
"""
no: int
prompt: Union[str, Any]
history_QA: Union[str, Any]
json_data: Optional[dict]
@@ -229,6 +229,7 @@ class EBDataset(BenchmarkDataset):
**kwargs,
) -> list:
samples: list = []
cnt = 1
for entry in self.data:
if len(samples) >= num_requests:
break
@@ -246,16 +247,17 @@ class EBDataset(BenchmarkDataset):
prompt, None)
samples.append(
SampleRequest(
no=cnt,
prompt=prompt,
prompt_len=self.prompt_len,
history_QA=[],
expected_output_len=new_output_len,
))
cnt += 1
self.maybe_oversample_requests(samples, num_requests)
return samples
class EBChatDataset(BenchmarkDataset):
"""
Implements the ShareGPT dataset. Loads data from a JSON file and generates
@@ -284,6 +286,7 @@ class EBChatDataset(BenchmarkDataset):
**kwargs,
) -> list:
samples: list = []
cnt = 1
for entry in self.data:
if len(samples) >= num_requests:
break
@@ -297,12 +300,14 @@ class EBChatDataset(BenchmarkDataset):
prompt, None)
samples.append(
SampleRequest(
no=cnt,
json_data=json_data,
prompt=prompt,
prompt_len=0,
history_QA=history_QA,
expected_output_len=new_output_len,
))
cnt += 1
self.maybe_oversample_requests(samples, num_requests)
return samples