apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-12-24 13:28:13 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Feature] Support for request-level speculative decoding metrics monitoring. (#5518)

Browse Source 
* support spec metrics monitor per request

* fix bug

* remove debug log

* fix ut bugs
This commit is contained in:
GoldPancake
2025-12-12 12:22:18 +08:00
committed by GitHub
parent 3c1f7b85a4
commit 909059c60a
11 changed files with 154 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/entrypoints/openai/test_completion_echo.py
View File

@@ -46,6 +46,7 @@ class TestCompletionEcho(unittest.IsolatedAsyncioTestCase):
"finished": True,
},
"output_token_ids": 3,
"metrics": {},
}
self.mock_engine.generate.return_value = [mock_output]
@@ -80,6 +81,7 @@ class TestCompletionEcho(unittest.IsolatedAsyncioTestCase):
"finished": True,
},
"output_token_ids": 3,
"metrics": {},
}
self.mock_engine.generate.return_value = [mock_output]
@@ -109,10 +111,12 @@ class TestCompletionEcho(unittest.IsolatedAsyncioTestCase):
{
"outputs": {"text": " response1", "token_ids": [1, 2], "top_logprobs": None, "finished": True},
"output_token_ids": 2,
"metrics": {},
},
{
"outputs": {"text": " response2", "token_ids": [3, 4], "top_logprobs": None, "finished": True},
"output_token_ids": 2,
"metrics": {},
},
]
self.mock_engine.generate.return_value = mock_outputs
@@ -146,10 +150,12 @@ class TestCompletionEcho(unittest.IsolatedAsyncioTestCase):
{
"outputs": {"text": " response1", "token_ids": [1, 2], "top_logprobs": None, "finished": True},
"output_token_ids": 2,
"metrics": {},
},
{
"outputs": {"text": " response2", "token_ids": [3, 4], "top_logprobs": None, "finished": True},
"output_token_ids": 2,
"metrics": {},
},
]
self.mock_engine.generate.return_value = mock_outputs

3
tests/entrypoints/openai/test_max_streaming_tokens.py
View File

@@ -321,6 +321,7 @@ class TestMaxStreamingResponseTokens(IsolatedAsyncioTestCase):
],
},
"finished": True,
"metrics": {},
},
{
"request_id": "test_request_id_1",
@@ -334,6 +335,7 @@ class TestMaxStreamingResponseTokens(IsolatedAsyncioTestCase):
],
},
"finished": True,
"metrics": {},
},
]
@@ -493,6 +495,7 @@ class TestMaxStreamingResponseTokens(IsolatedAsyncioTestCase):
prompt_logprobs_res_list=prompt_logprobs_res_list,
response_processor=mock_response_processor,
max_tokens=max_tokens_list[idx],
speculate_metrics=None,
)
expected = case["expected"]

2
tests/entrypoints/openai/test_serving_completion.py
View File

@@ -129,6 +129,7 @@ class TestOpenAIServingCompletion(unittest.IsolatedAsyncioTestCase):
"reasoning_token_num": 10,
},
"output_token_ids": 3,
"metrics": {},
},
{
"outputs": {
@@ -141,6 +142,7 @@ class TestOpenAIServingCompletion(unittest.IsolatedAsyncioTestCase):
"reasoning_token_num": 20,
},
"output_token_ids": 3,
"metrics": {},
},
]

9
tests/output/test_process_batch_output.py
View File

@@ -142,13 +142,10 @@ class TestTokenProcessorProcessBatchOutput(unittest.TestCase):
processor.num_accepted_tokens = 0
processor.num_emitted_tokens = 0
processor.max_num_emitted_tokens = 0
processor.num_rest_requests_per_head = [
0,
] * MAX_DRAFT_TOKENS
processor.num_accept_requests_per_head = [
0,
] * MAX_DRAFT_TOKENS
processor.speculative_stats_step = 0
processor.total_step_per_request = {}
processor.accept_token_num_per_head_per_request = {}
processor.accept_token_num_per_head = [0] * MAX_DRAFT_TOKENS
# processor._recycle_resources = Mock()