[Bugfix] Fix the issue where the qwen3-omni model long-term stability test sometimes gets stuck without sending requests. (#3468)
Signed-off-by: zhumingjue <zhumingjue@huawei.com>
This commit is contained in:
zhumingjue138 authored and GitHub committed
1 parent
dca369d448
commit
c167b9d691
5 files changed
+21
-83
No files matched your search
+17
-4
@@ -2,6 +2,7 @@ import json
|
||||
import os
|
||||
import re
|
||||
import subprocess
|
||||
import threading
|
||||
import time
|
||||
from datetime import datetime
|
||||
from pathlib import Path
|
||||
@@ -377,11 +378,23 @@ def run_benchmark(
|
||||
command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, bufsize=1, universal_newlines=True
|
||||
)
|
||||
|
||||
for line in iter(process.stdout.readline, ""):
|
||||
print(line, end=" ")
|
||||
if process.stdout is None or process.stderr is None:
|
||||
raise RuntimeError("Failed to capture benchmark process output streams")
|
||||
|
||||
for line in iter(process.stderr.readline, ""):
|
||||
print(line, end=" ")
|
||||
def _forward_stream(stream) -> None:
|
||||
try:
|
||||
for line in iter(stream.readline, ""):
|
||||
print(line, end="")
|
||||
finally:
|
||||
stream.close()
|
||||
|
||||
stdout_thread = threading.Thread(target=_forward_stream, args=(process.stdout,))
|
||||
stderr_thread = threading.Thread(target=_forward_stream, args=(process.stderr,))
|
||||
stdout_thread.start()
|
||||
stderr_thread.start()
|
||||
stdout_thread.join()
|
||||
stderr_thread.join()
|
||||
process.wait()
|
||||
|
||||
if "--result-dir" in command:
|
||||
index = command.index("--result-dir")
|
||||
|
||||
@@ -1,60 +1,8 @@
|
||||
[
|
||||
{
|
||||
"test_name": "test_qwen3_omni_stability",
|
||||
"server_params": {
|
||||
"model": "Qwen/Qwen3-Omni-30B-A3B-Instruct",
|
||||
"stage_overrides": {
|
||||
"2": {
|
||||
"max_num_batched_tokens": 1000000
|
||||
}
|
||||
}
|
||||
},
|
||||
"benchmark_params": [
|
||||
{
|
||||
"dataset_name": "random-mm",
|
||||
"backend": "openai-chat-omni",
|
||||
"endpoint": "/v1/chat/completions",
|
||||
"duration_sec": 86400,
|
||||
"request_rate": 0.3,
|
||||
"num_prompts_per_batch": 10,
|
||||
"random_input_len": {
|
||||
"min": 0,
|
||||
"max": 8000
|
||||
},
|
||||
"random_output_len": {
|
||||
"min": 0,
|
||||
"max": 1000
|
||||
},
|
||||
"random_range_ratio": 0.0,
|
||||
"random_mm_base_items_per_request": {
|
||||
"min": 0,
|
||||
"max": 6
|
||||
},
|
||||
"random_mm_num_mm_items_range_ratio": 0.0,
|
||||
"random_mm_limit_mm_per_prompt": {
|
||||
"image": 2,
|
||||
"video": 2,
|
||||
"audio": 2
|
||||
},
|
||||
"random_mm_bucket_config": {
|
||||
"(128-1024, 128-1024, 1)": 0.34,
|
||||
"(256-1080, 256-1920, 2-16)": 0.33,
|
||||
"(0, 1-60, 1-3)": 0.33
|
||||
},
|
||||
"ignore_eos": true,
|
||||
"percentile-metrics": "ttft,tpot,itl,e2el,audio_rtf,audio_ttfp,audio_duration"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"test_name": "test_qwen3_omni_stability_async_chunk",
|
||||
"server_params": {
|
||||
"model": "Qwen/Qwen3-Omni-30B-A3B-Instruct",
|
||||
"stage_overrides": {
|
||||
"2": {
|
||||
"max_num_batched_tokens": 1000000
|
||||
}
|
||||
},
|
||||
"extra_cli_args": ["--async-chunk"]
|
||||
},
|
||||
"benchmark_params": [
|
||||
@@ -64,7 +12,7 @@
|
||||
"endpoint": "/v1/chat/completions",
|
||||
"duration_sec": 86400,
|
||||
"max_concurrency": 2,
|
||||
"num_prompts_per_batch": 10,
|
||||
"num_prompts_per_batch": 50,
|
||||
"random_input_len": {
|
||||
"min": 0,
|
||||
"max": 8000
|
||||
|
||||
@@ -5,36 +5,13 @@
|
||||
"model": "Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice"
|
||||
},
|
||||
"benchmark_params": [
|
||||
{
|
||||
"dataset_name": "random",
|
||||
"backend": "openai-audio-speech",
|
||||
"endpoint": "/v1/audio/speech",
|
||||
"duration_sec": 86400,
|
||||
"request_rate": 0.3,
|
||||
"num_prompts_per_batch": 10,
|
||||
"random_input_len": {
|
||||
"min": 0,
|
||||
"max": 1000
|
||||
},
|
||||
"random_output_len": {
|
||||
"min": 0,
|
||||
"max": 1000
|
||||
},
|
||||
"random_range_ratio": 0.0,
|
||||
"extra_body": {
|
||||
"voice": "Vivian",
|
||||
"language": "English"
|
||||
},
|
||||
"ignore_eos": true,
|
||||
"percentile-metrics": "ttft,e2el,audio_rtf,audio_ttfp,audio_duration"
|
||||
},
|
||||
{
|
||||
"dataset_name": "random",
|
||||
"backend": "openai-audio-speech",
|
||||
"endpoint": "/v1/audio/speech",
|
||||
"duration_sec": 86400,
|
||||
"max_concurrency": 2,
|
||||
"num_prompts_per_batch": 10,
|
||||
"num_prompts_per_batch": 50,
|
||||
"random_input_len": {
|
||||
"min": 0,
|
||||
"max": 1000
|
||||
|
||||
@@ -11,7 +11,7 @@
|
||||
"backend": "vllm-omni",
|
||||
"duration_sec": 86400,
|
||||
"max_concurrency": 1,
|
||||
"num_prompts_per_batch": 10,
|
||||
"num_prompts_per_batch": 50,
|
||||
"width": {
|
||||
"min": 512,
|
||||
"max": 2048
|
||||
|
||||
@@ -19,7 +19,7 @@
|
||||
"backend": "v1/videos",
|
||||
"duration_sec": 86400,
|
||||
"max_concurrency": 1,
|
||||
"num_prompts_per_batch": 20,
|
||||
"num_prompts_per_batch": 50,
|
||||
"enable_negative_prompt": true,
|
||||
"random_request_config": [
|
||||
{"width": 832, "height": 480, "num_inference_steps": 4, "num_frames": 81, "fps": 16, "weight": 0.5},
|
||||
|
||||
Reference in new issue
Block a user