mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2026-04-23 00:17:25 +08:00
[Feature] console print statistical metrics (#6339)
* [Feature] console print statistical data * [Feature] console print statistical data v2 dp_rank * [Feature] console print statistical data v2 unittest * [Feature] console print statistical data v3 unittest
This commit is contained in:
@@ -146,6 +146,7 @@ class TestTokenProcessorProcessBatchOutput(unittest.TestCase):
|
||||
processor.engine_worker_queue = Mock()
|
||||
processor.split_connector = Mock()
|
||||
processor.resource_manager = MockResourceManager()
|
||||
processor.scheduler_metrics_logger = None
|
||||
task1 = MockTask()
|
||||
task2 = MockTask()
|
||||
processor.resource_manager.tasks_list = [task1, task2]
|
||||
|
||||
@@ -199,6 +199,21 @@ def test_init_allocates_expected_buffers():
|
||||
processor_logprob, _, _, _ = _make_processor(enable_logprob=True)
|
||||
assert list(processor_logprob.output_scores.shape) == [MAX_BSZ * (K + 1), 1]
|
||||
|
||||
|
||||
def test_is_decode_stage():
|
||||
processor, _, _, _ = _make_processor()
|
||||
|
||||
assert processor._is_decode_stage(None) is False
|
||||
|
||||
task = types.SimpleNamespace(need_prefill_tokens=None, num_computed_tokens=0)
|
||||
assert processor._is_decode_stage(task) is False
|
||||
|
||||
task = types.SimpleNamespace(need_prefill_tokens=4, num_computed_tokens=3)
|
||||
assert processor._is_decode_stage(task) is False
|
||||
|
||||
task = types.SimpleNamespace(need_prefill_tokens=4, num_computed_tokens=4)
|
||||
assert processor._is_decode_stage(task) is True
|
||||
|
||||
processor_spec, _, _, _ = _make_processor(speculative_method="mtp", enable_logprob=False)
|
||||
assert processor_spec.output_tokens.shape[0] == SPECULATE_MAX_BSZ * MAX_DRAFT_TOKENS + SPECULATE_MAX_BSZ + 2
|
||||
|
||||
|
||||
Reference in New Issue
Block a user