I am new to google cloud machine learning. I am setting up tensorflow training on the cloud and following the stock guide census data tutorial.
On my PC, I have python 3.6.7 (miniconda) with tensorflow 1.12. I have reached the point where the guide says to run training on local machine so I use gcloud command on my home PC as so:
gcloud ml-engine local train \
--module-name trainer.task \
--package-path trainer/ \
--job-dir $MODEL_DIR \
-- \
--train-files $TRAIN_DATA \
--eval-files $EVAL_DATA \
--train-steps 1000 \
--eval-steps 100
but get the error
/home/usman/miniconda3/bin/python: Error while finding module specification for 'trainer.task' (ModuleNotFoundError: No module named 'trainer')
Output of gcloud info on local machine:
Cloud SDK on PATH: [False]
Kubectl on PATH: [False]
Installation Properties: [/usr/lib/google-cloud-sdk/properties]
User Config Directory: [/home/usman/.config/gcloud]
Active Configuration Name: [default]
Active Configuration Path: [/home/usman/.config/gcloud/configurations/config_default]
Account: [drusmanbashir@gmail.com]
Project: [api-project-988931422863]
Current Properties:
[core]
project: [api-project-988931422863]
account: [drusmanbashir@gmail.com]
disable_usage_reporting: [True]
[compute]
region: [europe-west2]
zone: [europe-west2-b]
Logs Directory: [/home/usman/.config/gcloud/logs]
Last Log File: [/home/usman/.config/gcloud/logs/2018.12.03/13.18.07.624045.log]
git: [git version 2.17.1]
ssh: [OpenSSH_7.6p1 Ubuntu-4ubuntu0.1, OpenSSL 1.0.2n 7 Dec 2017]
I am sure the problem is with setting up env or something, will be grateful for the advice.