Bert训练模型时报错：Windows fatal exception: access violation

运行之后：

2020-04-28 23:26:44.729208: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
Windows fatal exception: access violation

Current thread 0x00004a90 (most recent call first):
  File "C:\Users\DELL\Anaconda3\envs\tensorflow-gpu\lib\site-packages\tensorflow_core\python\lib\io\file_io.py", line 84 in _preread_check
  File "C:\Users\DELL\Anaconda3\envs\tensorflow-gpu\lib\site-packages\tensorflow_core\python\lib\io\file_io.py", line 122 in read
      File "D:\senti\code\Bert\modeling.py", line 94 in from_json_file
  File "D:/senti/code/Bert/run_classifier.py", line 844 in main
  File "C:\Users\DELL\Anaconda3\envs\tensorflow-gpu\lib\site-packages\absl\app.py", line 250 in _run_main
  File "C:\Users\DELL\Anaconda3\envs\tensorflow-gpu\lib\site-packages\absl\app.py", line 299 in run
  File "C:\Users\DELL\Anaconda3\envs\tensorflow-gpu\lib\site-packages\tensorflow_core\python\platform\app.py", line 40 in run
  File "D:/senti/code/Bert/run_classifier.py", line 1025 in <module>

Process finished with exit code -1073741819 (0xC0000005)

降低一下tf版本

还有一点是确定使用的GPU

你解决了吗

我跟你遇到了一样的错误，想问大家都解决了吗，怎么弄得？？？

建议检查一下config，特别是文件读取的路径有没有问题。可以新建一个.py文件尝试一下。

import json

with open('./GLUE/BERT_MODEL_DIR/uncased_L-12_H-768_A-12/bert_config.json', 'r') as file:
    data = json.load(file)