81f3c09d6d
### What this PR does / why we need it? This PR updates the CI runner from `linux-aarch64-a2-*` to `linux-aarch64-a2b3-*` in various test configuration files. This change is necessary to adapt to updates in the CI infrastructure. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? The changes are configuration updates for CI tests. The correctness will be verified by the CI pipeline. Signed-off-by: hfadzxy <starmoon_zhang@163.com>
112 lines
3.0 KiB
Python
112 lines
3.0 KiB
Python
import json
|
|
import os
|
|
from typing import Any
|
|
|
|
import openai
|
|
import pytest
|
|
from vllm.utils.network_utils import get_open_port
|
|
|
|
from tests.e2e.conftest import RemoteOpenAIServer
|
|
from tools.aisbench import run_aisbench_cases
|
|
|
|
MODELS = [
|
|
"Qwen/Qwen3-Next-80B-A3B-Instruct",
|
|
]
|
|
|
|
MODES = ["aclgraph"]
|
|
|
|
TENSOR_PARALLELS = [4]
|
|
MAX_NUM_BATCHED_TOKENS = [8192, 32768]
|
|
|
|
prompts = [
|
|
"San Francisco is a",
|
|
]
|
|
|
|
api_keyword_args = {
|
|
"max_tokens": 10,
|
|
}
|
|
|
|
batch_size_dict = {
|
|
"linux-aarch64-a2b3-4": 64,
|
|
"linux-aarch64-a3-4": 64,
|
|
}
|
|
VLLM_CI_RUNNER = os.getenv("VLLM_CI_RUNNER", "linux-aarch64-a2b3-4")
|
|
performance_batch_size = batch_size_dict.get(VLLM_CI_RUNNER, 1)
|
|
|
|
aisbench_cases = [{
|
|
"case_type": "performance",
|
|
"dataset_path": "vllm-ascend/GSM8K-in3500-bs400",
|
|
"request_conf": "vllm_api_stream_chat",
|
|
"dataset_conf": "gsm8k/gsm8k_gen_0_shot_cot_str_perf",
|
|
"num_prompts": 4 * performance_batch_size,
|
|
"max_out_len": 1500,
|
|
"batch_size": performance_batch_size,
|
|
"baseline": 1,
|
|
"threshold": 0.97
|
|
}, {
|
|
"case_type": "accuracy",
|
|
"dataset_path": "vllm-ascend/gsm8k-lite",
|
|
"request_conf": "vllm_api_general_chat",
|
|
"dataset_conf": "gsm8k/gsm8k_gen_0_shot_cot_chat_prompt",
|
|
"max_out_len": 32768,
|
|
"batch_size": 32,
|
|
"top_k": 20,
|
|
"baseline": 95,
|
|
"threshold": 5
|
|
}]
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.parametrize("model", MODELS)
|
|
@pytest.mark.parametrize("mode", MODES)
|
|
@pytest.mark.parametrize("tp_size", TENSOR_PARALLELS)
|
|
@pytest.mark.parametrize("max_num_batched_tokens", MAX_NUM_BATCHED_TOKENS)
|
|
async def test_models(model: str, mode: str, tp_size: int,
|
|
max_num_batched_tokens: int) -> None:
|
|
port = get_open_port()
|
|
env_dict = {
|
|
"OMP_NUM_THREADS": "10",
|
|
"OMP_PROC_BIND": "false",
|
|
"HCCL_BUFFSIZE": "1024",
|
|
"PYTORCH_NPU_ALLOC_CONF": "expandable_segments:True",
|
|
}
|
|
server_args = [
|
|
"--tensor-parallel-size",
|
|
str(tp_size),
|
|
"--port",
|
|
str(port),
|
|
"--max-model-len",
|
|
"40960",
|
|
"--max-num-batched-tokens",
|
|
str(max_num_batched_tokens),
|
|
"--trust-remote-code",
|
|
"--async-scheduling",
|
|
"--no-enable-prefix-caching",
|
|
"--enable-expert-parallel",
|
|
"--gpu-memory-utilization",
|
|
"0.8",
|
|
"--max-num-seqs",
|
|
"64",
|
|
]
|
|
request_keyword_args: dict[str, Any] = {
|
|
**api_keyword_args,
|
|
}
|
|
with RemoteOpenAIServer(model,
|
|
server_args,
|
|
server_port=port,
|
|
env_dict=env_dict,
|
|
auto_port=False) as server:
|
|
client = server.get_async_client()
|
|
batch = await client.completions.create(
|
|
model=model,
|
|
prompt=prompts,
|
|
**request_keyword_args,
|
|
)
|
|
choices: list[openai.types.CompletionChoice] = batch.choices
|
|
assert choices[0].text, "empty response"
|
|
print(choices)
|
|
if mode == "single":
|
|
return
|
|
# aisbench test
|
|
run_aisbench_cases(model, port, aisbench_cases)
|