使用Kunlunxin XPU通过FastDeploy部署模型错误

[dddlli]

XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded
[12:30:35][jupyter-338262-9442594][125547:125547][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35
/opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/_distutils_hack/__init__.py:30: UserWarning: Setuptools is replacing distutils. Support for replacing an already imported distutils is deprecated. In the future, this condition will fail. Register concerns at https://github.com/pypa/setuptools/issues/new?template=distutils-deprecation.yml
  warnings.warn(
[2025-08-27 12:30:37,233] [    INFO] distributed_strategy.py:333 - distributed strategy initialized
======================= Modified FLAGS detected =======================
FLAGS(name='FLAGS_selected_xpus', current_value='0', default_value='')
FLAGS(name='FLAGS_enable_pir_in_executor', current_value=True, default_value=False)
FLAGS(name='FLAGS_specialize_device_in_dy2st', current_value=True, default_value=False)
FLAGS(name='FLAGS_pir_interpreter_record_stream_for_gc_cache', current_value=True, default_value=False)
FLAGS(name='FLAGS_parameters_persistent_mode_in_dy2st', current_value=True, default_value=False)
=======================================================================
/opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/distributed/parallel.py:1054: UserWarning: Currently not a parallel execution environment, `paddle.distributed.init_parallel_env` will not do anything.
  warnings.warn(
[2025-08-27 12:30:37,235] [    INFO] topology.py:525 - Total 1 pipe comm group(s) create successfully!
/opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/distributed/communication/group.py:145: UserWarning: Current global rank 0 is not in group _default_pg10
  warnings.warn(
[2025-08-27 12:30:37,243] [    INFO] topology.py:525 - Total 1 data comm group(s) create successfully!
[2025-08-27 12:30:37,243] [    INFO] topology.py:525 - Total 1 model comm group(s) create successfully!
[2025-08-27 12:30:37,244] [    INFO] topology.py:525 - Total 1 sharding comm group(s) create successfully!
[2025-08-27 12:30:37,244] [    INFO] topology.py:439 - HybridParallelInfo: rank_id: 0, mp_degree: 1, sharding_degree: 1, pp_degree: 1, dp_degree: 1, sep_degree: 1, mp_group: [0],  sharding_group: [0], pp_group: [0], dp_group: [0], sep:group: None, check/clip group: [0]
�[32m[2025-08-27 12:30:37,244] [    INFO]�[0m - Loading configuration file baidu/ERNIE-4.5-21B-A3B-Paddle/config.json�[0m
�[33m[2025-08-27 12:30:37,244] [ WARNING]�[0m - You are using a model of type ernie4_5_moe to instantiate a model of type . This is not supported for all configurations of models and can yield errors.�[0m
Traceback (most recent call last):
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/engine/../worker/worker_process.py", line 753, in <module>
    run_worker_proc()
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/engine/../worker/worker_process.py", line 728, in run_worker_proc
    worker_proc = PaddleDisWorkerProc(fd_config, ranks, local_rank)
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/engine/../worker/worker_process.py", line 149, in __init__
    self.worker = get_worker(fd_config=fd_config, local_rank=self.local_rank, rank=self.ranks)
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/engine/../worker/worker_process.py", line 67, in get_worker
    from fastdeploy.worker.xpu_worker import XpuWorker
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/worker/xpu_worker.py", line 29, in <module>
    from fastdeploy.worker.xpu_model_runner import XPUModelRunner
  File "/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/worker/xpu_model_runner.py", line 41, in <module>
    from fastdeploy.model_executor.ops.xpu import (
ImportError: cannot import name 'adjust_batch' from 'fastdeploy.model_executor.ops.xpu' (/home/aistudio/external-libraries/lib/python3.10/site-packages/fastdeploy/model_executor/ops/xpu/__init__.py)

[hong19860320]

先跑一下这些命令吧，可以把日志发上来

python -c "import paddle; paddle.version.show()"
python -c "import paddle; paddle.utils.run_check()"
python -c "from paddle.jit.marker import unified"
python -c "from fastdeploy.model_executor.ops.xpu import block_attn"

[dddlli]

XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded
[11:36:54][jupyter-338262-9442594][1343:1343][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35
full_version: 3.1.1
major: 3
minor: 1
patch: 1
rc: 0
cuda: False
cudnn: False
nccl: 0
xpu_xre: 5.0.21.26
xpu_xccl: 3.0.2.7
xpu_xhpc: dev/20990602
cinn: False
tensorrt: None
cuda_archs: []
XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded
[11:36:55][jupyter-338262-9442594][1532:1532][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35
Running verify PaddlePaddle program ...
I0828 11:36:56.412321 1532 pir_interpreter.cc:1524] New Executor is Running ...
W0828 11:36:56.412358 1532 xpu_context.cc:186] Please NOTE: xpu device: 0
I0828 11:36:56.412945 1532 pir_interpreter.cc:1547] pir interpreter is running by multi-thread mode ...
PaddlePaddle works well on 1 XPU.
PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.
XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded
[11:36:56][jupyter-338262-9442594][1760:1760][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35
XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded
[11:36:58][jupyter-338262-9442594][1942:1942][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35
/opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/_distutils_hack/init.py:30: UserWarning: Setuptools is replacing distutils. Support for replacing an already imported distutils is deprecated. In the future, this condition will fail. Register concerns at https://github.com/pypa/setuptools/issues/new?template=distutils-deprecation.yml
warnings.warn(

[hong19860320]

XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded [11:36:54][jupyter-338262-9442594][1343:1343][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35 full_version: 3.1.1 major: 3 minor: 1 patch: 1 rc: 0 cuda: False cudnn: False nccl: 0 xpu_xre: 5.0.21.26 xpu_xccl: 3.0.2.7 xpu_xhpc: dev/20990602 cinn: False tensorrt: None cuda_archs: [] XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded [11:36:55][jupyter-338262-9442594][1532:1532][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35 Running verify PaddlePaddle program ... I0828 11:36:56.412321 1532 pir_interpreter.cc:1524] New Executor is Running ... W0828 11:36:56.412358 1532 xpu_context.cc:186] Please NOTE: xpu device: 0 I0828 11:36:56.412945 1532 pir_interpreter.cc:1547] pir interpreter is running by multi-thread mode ... PaddlePaddle works well on 1 XPU. PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now. XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded [11:36:56][jupyter-338262-9442594][1760:1760][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35 XCCL /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/paddle/libs/libbkcl.so loaded [11:36:58][jupyter-338262-9442594][1942:1942][WARN][BKCL][globals.cpp:259] xccl version: 7b9c31e [rdma] [fix compile issue] build data: May 13 2025 06:07:35 /opt/conda/envs/python35-paddle120-env/lib/python3.10/site-packages/_distutils_hack/init.py:30: UserWarning: Setuptools is replacing distutils. Support for replacing an already imported distutils is deprecated. In the future, this condition will fail. Register concerns at https://github.com/pypa/setuptools/issues/new?template=distutils-deprecation.yml warnings.warn(

有完整的吗？我看是不是还有日志呢？

[hong19860320]

另外，你有直接基于我们提供的最新镜像测试吗？https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/get_started/installation/kunlunxin_xpu.md

mkdir Work
cd Work
docker pull ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.1.0
docker run --name fastdeploy-xpu --net=host -itd --privileged -v $PWD:/Work -w /Work \
    ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.1.0 \
    /bin/bash
docker exec -it fastdeploy-xpu /bin/bash