search

google-research / electra

ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
Apache License 2.0
2.31k stars 351 forks source link

"Device or resource busy" for mounted paths #78

Open emirkin opened 4 years ago

emirkin commented 4 years ago

Running

python3 build_openwebtext_pretraining_dataset.py --data-dir data --num-processes 8

gives the error:

Traceback (most recent call last):
  File "build_openwebtext_pretraining_dataset.py", line 103, in <module>
    main()
  File "build_openwebtext_pretraining_dataset.py", line 89, in main
    utils.rmkdir(os.path.join(args.data_dir, "pretrain_tfrecords"))
  File "/home/zm/electra/util/utils.py", line 64, in rmkdir
    rmrf(path)
  File "/home/zm/electra/util/utils.py", line 60, in rmrf
    tf.io.gfile.rmtree(path)
  File "/home/zm/anaconda3/envs/electra/lib/python3.7/site-packages/tensorflow_core/python/lib/io/file_io.py", line 569, in delete_recursively_v2
    pywrap_tensorflow.DeleteRecursively(compat.as_bytes(path))
tensorflow.python.framework.errors_impl.FailedPreconditionError: data/pretrain_tfrecords; Device or resource busy

My data/pretrain_tfrecords is a mounted path.