👉 遇到问题先看这里

🌟 第一次使用工具遇到问题？

按🚀 快速入门

❓ 搜索历史issue，查看冷门的同类问题

在🔖 Issue中搜索历史类似问题

操作系统及版本

openEuler 22.03 (LTS-SP4)

python版本

3.11

MsModelSlim 工具版本

84b2f65060fff6a91ad7f1858c92bcede83f9d11

是否必现

是

MsModelSlim 执行命令

msmodelslim tune --model_path ${MODEL_PATH} --save_path ${SAVE_PATH} --config qwen3_32b_w8a8_standing_high_gsm8k.yaml --device npu:4,5,6,7 --model_type Qwen3-32B --trust_remote_code True

预期行为

1、设置浮点精度目标为100，首轮执行浮点精度测试，不达预期则会报错退出

实际行为

1、量化精度自动调优报错退出，没有正常进入量化和评测流程
2026-09-05 16:51:49,789 - msmodelslim.core.runner.layer wise runner - INFO - Run processor BinaryOperatorLayerWiseProcessor for "model.lavers.60"a
2026-09-05 16:51:54.675 msmodelslim.core.runner.layer wise_runner INFO Run processor LoadProcessor(device=meta, non blocking=False) for "model.layers.60"er
2026-09-05 16:51:55.450 msmodelslim.core.runner.layer wise_runner INFO Run processor LoadProcessor(device=npu:O, non_blocking=False) for "model.layers.61"cc
2026-09-05 16:51:55,925 msmodelslim.core.runner.layer_wise_runner INFO Run processor BinaryOperatorLayerWiseProcessor for "model.layers.61"a
2026-09-05 16:52:01.237 msmodelslim.core.runner.layer wise runner INFO Run processor LoadProcessor(device=meta, non blocking=False) for "model.layers.61"
2026-09-05 16:52:01,899 msmodelslim.core.runner.layer wise_runner INFO的 Run processor LoadProcessor(device=npu:0, non blocking=False) for "model.layers.62"
2026-09-05 16:52:02.497 msmodelslim.core.runner.layer wise_runner INFO Run processor BinaryOperatorLayerWiseProcessor for "model.lavers.62"a
2026-09-05 16:52:07.570 msmodelslim.core.runner.layer wise_runner INFO Run processor LoadProcessor(device=meta, non blocking=False) for "model.layers.62"
2026-09-05 16:52:08.286 msmodelslim.core.runner.layer wise runner INFO - Run processor LoadProcessor(device=npu:0, non blocking=False) for "model.lavers.63"
2026-09-05 16:52:08,755 msmodelslim.core.runner.layer wise runner INFO - Run processor BinaryOperatorLayerWiseProcessor for "model.lavers.63"
2026-09-05 16:52:13,846 msmodelslim.core.runner.layer wise_runner INFO - Run processor LoadProcessor(device=meta, non blocking=False) for "model.layers.63"
2026-09-05 16:52:14.545 msmodelslim.app.auto tuning.application INFO - BinaryOperatorLayerWiseProcessor post run: 64 layer scores (mse layer wise), quant modules=['*']
2026-09-05 16:52:14.838 msmodelslim.app.auto tuning.application INFO - Generate practice success
Traceback (most recent call last):
File "/usr/local/python3.11.10/bin/msmodelslim", line 8, in
sys.exit(main())
^^^^AA
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/cli/ main_.py", line 1059, in main
tuning main argse
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/cli/auto_tuning/main.py", line 99, in main
app.tune(
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/utils/logging.py", line 314, in wrapper
return func(*args, **kwargs)
VVVvwVvwwvwywyvwvwwwv
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/app/auto tuning/application.py", Line 109, in tune
self. tune(model type, model path, save path, plan id, device, device indices, timeout,
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/utils/logging.py", line 314, in wrapper
return func(*args, **kwargs)
wrvyvvvvyyyyyveyvywvy
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/app/auto tuning/application.py", line 196, in _tune
self.quantization service.quantize(
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/utils/logging.py", line 314, in wrapper
return func(*args, **kwargs )
VVvwvvwyvywyvwevwwvwy
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/core/quant service/proxy/quant service.py", line 90, in quantize
quant service.quantize( c
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/utils/logging.py", line 314, in wrapper
return func(*args, **kwargs)
vVyVvvvvwvwwwwwwwwwww
File "/usr/local/python3.11.10/lib/python3.11/site-packages/msmodelslim/core/quant service/modelslim_v1/quant service.py", line 134, in quantize
raise SchemaValidateError(
msmodelslim.utils.exception.SchemaValidateError: Code: 203, Message: task is NOT BaseQuantConfig
TIP: Please make sure the task is a BaseQuantConfig
[ERROR] 2026-09-05-16:52:15 (PID:4285, Device:0, RankID:-1) ERR99999 UNKNOWN applicaiton exception

欢迎加入社区，感谢您对社区的贡献 🎉!