search

ChrisCummins / phd

👨‍💻 My PhD.
http://chriscummins.cc/
185 stars 34 forks source link

Got An Error on ‘docker run chriscummins/opencl_fuzz’ #25

Open Implementist opened 6 years ago

Implementist commented 6 years ago

Hi，Mr.Cummins When I try to do docker run chriscummins/opencl_fuzz I got an error like: I0728 03:12:19.383133 140557201053440 opencl_fuzz.py:331] Recording interesting results in /out. I0728 03:12:19.383535 140557201053440 opencl_fuzz.py:333] Preparing generator. Traceback (most recent call last): File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz.py", line 373, in <module> app.run(main) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/pypi__absl_py_0_1_10/absl/app.py", line 274, in run _run_main(main, argv) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/pypi__absl_py_0_1_10/absl/app.py", line 238, in _run_main sys.exit(main(argv)) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz.py", line 336, in main clgen_pretrained.ClgenGenerator) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz.py", line 311, in GeneratorFromFlag return generator_class(config) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/deeplearning/deepsmith/generators/clgen_pretrained.py", line 53, in __init__ self.instance = sample.Instance(self.config.instance) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/deeplearning/clgen/sample.py", line 46, in __init__ pathlib.Path(config.pretrained_model)) File "/app/experimental/deeplearning/deepsmith/opencl_fuzz/opencl_fuzz_image.binary.runfiles/phd/deeplearning/clgen/models/pretrained.py", line 31, in __init__ self.cache = cache.FSCache(self.path) AttributeError: module 'deeplearning.clgen.cache' has no attribute 'FSCache'

How can I fix this?

ChrisCummins commented 6 years ago

Hi @Implementist , thanks for the bug report! I see the root cause and will push a fix shortly. In the mean time, could you help me reproduce the issue - what operating system are you using, and which version of docker?

Cheers, Chris

Implementist commented 6 years ago

Thanks for your reply. I am using CentOS 7 and docker version is 1.13.1.

Implementist commented 6 years ago

Hi Mr.Cummins, Today I tried that command on another Server with Ubuntu 16.04 and docker '18.06.0-ce, build 0ffa825'. And I got the same error.

ChrisCummins commented 6 years ago

Hi @Implementist , thanks for the further details. Confusingly, I have been unable to reproduce the error on any of my machines (I'm no Docker expert so maybe I'm making a simple mistake, but I thought the advantage of containerized apps is that it makes these sorts of problems reproducible on any system?).

Nevertheless, I have pushed a new docker image which contains an (untested) fix, sha256:ef58b41d0a63e79923bb7bc0bb834cb397ae9a05e7bcc9fa2122af9f5a303ce2.

If you update your image with docker pull chriscummins/opencl_fuzz the problem should go away.

Cheers, Chris

ChrisCummins commented 5 years ago

Hi @Implementist, were you able to test my fix for this issue?