fashion-mnist-fairing 예제 실행 오류 문의

[nuri428]

예제로 있는 fashion-mnist-fairing.ipynb 파일을 실행 시키면 아래와 같은 오류가 발생 합니다.

[I 210105 19:06:39 config:134] Using preprocessor: <kubeflow.fairing.preprocessors.converted_notebook.ConvertNotebookPreprocessor object at 0x7f752045bbe0>
[I 210105 19:06:39 config:136] Using builder: <kubeflow.fairing.builders.append.append.AppendBuilder object at 0x7f74800fa198>
[I 210105 19:06:39 config:138] Using deployer: <kubeflow.fairing.deployers.job.job.Job object at 0x7f75203ffb70>
[W 210105 19:06:39 append:50] Building image using Append builder...
[I 210105 19:06:39 base:107] Creating docker context: /tmp/fairing_context_v1cjs1yb

TypeError Traceback (most recent call last)
in
43
44 )
---> 45 fairing.config.run()
46 else:
47 remote_train = MyFashionMnist()

~/.local/lib/python3.6/site-packages/kubeflow/fairing/config.py in run(self)
138 logging.info("Using deployer: %s", deployer)
139
--> 140 builder.build()
141 pod_spec = builder.generate_pod_spec()
142 deployer.deploy(pod_spec)

~/.local/lib/python3.6/site-packages/kubeflow/fairing/builders/append/append.py in build(self)
50 logger.warning("Building image using Append builder...")
51 start = timer()
---> 52 new_img = self._build(transport, src)
53 end = timer()
54 logger.warning("Image successfully built in {}s.".format(end-start))

~/.local/lib/python3.6/site-packages/kubeflow/fairing/builders/append/append.py in _build(self, transport, src)
64
65 def _build(self, transport, src):
---> 66 file, hash = self.preprocessor.context_tar_gz() # pylint:disable=redefined-builtin
67 self.context_file, self.context_hash = file, hash
68 self.image_tag = self.full_image_name(self.context_hash)

~/.local/lib/python3.6/site-packages/kubeflow/fairing/preprocessors/base.py in context_tar_gz(self, output_file)
106 , output_file = tempfile.mkstemp(prefix="/tmp/fairing_context")
107 logging.info("Creating docker context: %s", output_file)
--> 108 self.input_files = self.preprocess()
109 with tarfile.open(output_file, "w:gz", dereference=True) as tar:
110 for dst, src in self.context_map().items():

~/.local/lib/python3.6/site-packages/kubeflow/fairing/preprocessors/converted_notebook.py in preprocess(self)
119 exporter = nbconvert.PythonExporter()
120 exporter.register_preprocessor(self.notebook_preprocessor, enabled=True)
--> 121 contents, _ = exporter.from_filename(self.notebook_file)
122 converted_notebook = Path(self.notebook_file).with_suffix('.py')
123 if converted_notebook.exists() and not self.overwrite:

/usr/local/lib/python3.6/dist-packages/nbconvert/exporters/exporter.py in from_filename(self, filename, resources, **kw)
165 if not 'metadata' in resources or resources['metadata'] == '':
166 resources['metadata'] = ResourcesDict()
--> 167 path, basename = os.path.split(filename)
168 notebook_name = os.path.splitext(basename)[0]
169 resources['metadata']['name'] = notebook_name

/usr/lib/python3.6/posixpath.py in split(p)
105 """Split a pathname. Returns tuple "(head, tail)" where "tail" is
106 everything after the final slash. Either part may be empty."""
--> 107 p = os.fspath(p)
108 sep = _get_sep(p)
109 i = p.rfind(sep) + 1

TypeError: expected str, bytes or os.PathLike object, not NoneType

해당 내용에 대해서 문의 드립니다.

[nuri428]

해당 내용은 제가 Jupyter Lab을 이용해서 발생한 문제네요
전처리 모듈이 JupyterLab을 통해서 ipynb 파일을 가져오지 못하고 있고 다른 방법이 없어서
코드를 일부 수정 해야합니다.

fairing.config.set_builder(
'append',
image_name='fairing-job',
base_image='brightfly/kubeflow-jupyter-lab:tf2.0-gpu',
registry=DOCKER_REGISTRY,
push=True)

    fairing.config.set_preprocessor(name='notebook', notebook_file='test.ipynb')
    # 원래 예제 코드에서 아래에 이 코드를 추가 하면 됩니다. 
    # name :   전처리기 클래스명(주피터 노트북일 경우 notebook)
    # notebook_file : ipynb 파일명

    # cpu 2, memory 5GiB
    fairing.config.set_deployer('job',
                                namespace='dudaji',
                                pod_spec_mutators=[
                                    k8s_utils.get_resource_mutator(cpu=2,
                                                                   memory=5)]
     
                               )