1

Why do I get a subprocess object when I excute the code locally with gcloud? How can I see the execution of my training code?

$ gcloud ai-platform local train --module-name trainer.final_task --package-path trainer/                                                                                              

WARNING: Unexpected tensorflow version /Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint8 = np.dtype([("qint8", np.int8, 1)])
/Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint16 = np.dtype([("qint16", np.int16, 1)])
/Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint32 = np.dtype([("qint32", np.int32, 1)])
/Users/antonio/anaconda3/envs/ml6/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  np_resource = np.dtype([("resource", np.ubyte, 1)])
1.13.1
, using the default primary node name, aka "chief" for cluster settings
<subprocess.Popen object at 0x10e4b5590>
polodealvarado
  • 11
  • 2
  • What are your requirements.txt? All of the issues here seem to be version related issues. The warning is related to incompatibilities between your TF and numpy versions (see https://stackoverflow.com/questions/57381430/synonym-of-type-is-deprecated-in-a-future-version-of-numpy-it-will-be-underst). Can you try adding [import warnings warnings.filterwarnings('ignore')?]. However, the subprocess issue seems to be independent from the warning. Can you check the issue reported in https://github.com/GoogleCloudPlatform/cloudml-samples/issues/476? Are there any similarities with your training? – Albert Albesa Mar 23 '20 at 10:10

0 Answers0