bug描述 Describe the Bug
使用的paddle版本:
paddlepaddle-gpu 1.6.2.post107
cuda版本:
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:33:58_PDT_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0
报错信息:
W0620 12:53:27.994474 149242 device_context.cc:236] Please NOTE: device: 0, CUDA Capability: 80, Driver API Version: 12.0, Runtime API Version: 10.0
W0620 12:53:27.995631 149242 device_context.cc:244] device: 0, cuDNN Version: 8.6.
W0620 12:53:28.540676 149242 operator.cc:179] truncated_gaussian_random raises an exception thrust::system::system_error, parallel_for failed: no kernel image is available for execution on the device
/root/paddlejob/workspace/env_run/valuation/python2/lib/python2.7/site-packages/paddle/fluid/executor.py:779: UserWarning: The following exception is not an EOF exception.
"The following exception is not an EOF exception.")
Traceback (most recent call last):
File "run_with_json.py", line 112, in
trainer = build_trainer(trainer_params_dict, dataset_reader, model, num_train_examples)
File "run_with_json.py", line 83, in build_trainer
trainer = trainer_class(params=params_dict, data_set_reader=dataset_reader, model_class=model)
File "/root/paddlejob/workspace/env_run/valuation/textone/training/custom_trainer.py", line 30, in init
BaseTrainer.init(self, params, data_set_reader, model_class)
File "/root/paddlejob/workspace/env_run/valuation/textone/training/base_trainer.py", line 53, in init
self.executor.run(self.startup_program)
File "/root/paddlejob/workspace/env_run/valuation/python2/lib/python2.7/site-packages/paddle/fluid/executor.py", line 780, in run
six.reraise(*sys.exc_info())
File "/root/paddlejob/workspace/env_run/valuation/python2/lib/python2.7/site-packages/paddle/fluid/executor.py", line 775, in run
use_program_cache=use_program_cache)
File "/root/paddlejob/workspace/env_run/valuation/python2/lib/python2.7/site-packages/paddle/fluid/executor.py", line 822, in _run_impl
use_program_cache=use_program_cache)
File "/root/paddlejob/workspace/env_run/valuation/python2/lib/python2.7/site-packages/paddle/fluid/executor.py", line 899, in _run_program
fetch_var_name)
RuntimeError: parallel_for failed: no kernel image is available for execution on the device
其他补充信息 Additional Supplementary Information
No response
1条答案
按热度按时间k7fdbhmy1#
您好,从使用的paddle版本paddlepaddle-gpu 1.6.2.post107来看是基于cuda10.7的版本,与cuda11.8环境不符合,需升级paddle版本,参考官网paddle安装说明
python -m pip install paddlepaddle-gpu==2.6.1 -i https://pypi.tuna.tsinghua.edu.cn/simple