Quick route for linux command line

(A somewhat critical look at Google’s materials)

I was asked to do a workshop for a Google event in Thailand. Normally, my choice would be to prepare something new and interesting for an audience, and have had some success in the past…

However, the Google team was keen that instead I should do a run-through of their pre-made BabyWeights Qwiklab - despite the fact that I expressed reservations :

  • I had previously done the lab before, and hardly remember anything about it. This is a negative

  • What I did remember was more about the frustrations, than the cool factor

This quick write-up is just an aide-memoire for me, so that I can go through on Saturday without messing up too much. We’ll see how critical it turns out. Cross-fingers.

Start from scratch

  • From the email :

    “Thanks for volunteering to teach the ‘Baby-weight prediction’ tutorial in Qwiklabs on Sat, July 8th. For you to prepare, please use the following access code : bd35-dead-beef-2015

  • Following the email link :

    • New browser window : “Sorry, access denied to this resource.”
    • Sign into Qwiklabs : Wants an email or a Google Login (chose Google)
    • Must accept the 4000+ word “Qwiklabs Terms of Service” (I Accept)

    • “You don’t have any history yet! Enroll in an instructor-led course, get started on a self-paced course, or take a short lab to get started.”
    • My Account - Credits & Subscriptions - “Buy Credits or Subscriptions” button
    • Tiny link near bottom : “Have Promo Code?” (enter ‘access’ code given - hopefully it’s a ‘promo’ code)
    • FAILURE

    • Look at See how to apply a Qwiklab access code : 1 Page Presentation
    • “go to the Course or the codelab you wish to use” : (Search for “baby” in the Search Bar)
    • Pick third link shown (only one with “baby” in the title)
    • “Start Lab”-button -> “Enter Lab Access Code” (Cannot copy-paste code)
      • Type in code from email : Press button
      • Button didn’t seem to work, press it again
    • QwikLab counter seems to start to go down, but there’s a big red box at the top of the page :
      • “Sorry, this token was already used by Martin on Wed, 04 Jul 2018 11:05:33 -0400” = Sigh, pressing on..?
  • Presentation now says : “Qwiklabs generates a new temporary Google Account for you”
    • Is this the stuff on the Left Hand side?
    • As instructed, open a new Incognito Window, and go to the Google Cloud Console there
    • Agree to the terms of Service
    • But it already chose one of my existing accounts, not the new, temporary. one
      • Sign Out (?)
    • Account = Pull-down to “Use another account”
      • Use (randomised) username and password : “Welcome to your new account”
      • Accept large T&Cs paragraph
      • “Protect your account” : nope, don’t need. “Done”
    • Agree to the terms of Service (for temporary account) : “Agree and Continue”
    • New account creation process took 10mins of the QwikLabs time
  • Now seem to be signed in :
    • Use the Hamburger in Top Left to shrink ‘Catalog’ sidebar
    • Use little arrow on Bottom Left to shrink ‘Connection Details’ sidebar
    • AHAH! Now the QwikLabs thing fits on my laptop screen at regular 100% size

    • Finally, on to the Lab itself…
      • Why does the Lab timer keep jumping down? My machine is not under any load
    • Instructions now state :
      • “If you already have your own GCP account, make sure you do not use it for this lab.”
    • Ohhh : Now I see that the instructions are actually on this page, just several page-downs away

Launch Cloud Datalab

  • In the Incognito (new user) window, use the leftmost tiny icon on the top right to:
    • Open a dialog box with an animation, which seems to be typing stuff into a shell for me
    • Decide that it’s “fake”, and press “Start Cloud Shell” highlighted text at the bottom
  • Apparently, I should type in the text in the black boxes :
    • gcloud auth list : Seems to produce output
$ gcloud auth list
          Credentialed Accounts
ACTIVE  ACCOUNT
*       google646468_student@qwiklabs.net
To set the active account, run:
    $ gcloud config set account `ACCOUNT`

The account does seem to be active, but looks different from the instructions. Puzzling.

$ gcloud config list project
ERROR: (gcloud.config.list) The project property is set to the empty string, which is invalid.
To set your project, run:
  $ gcloud config set project PROJECT_ID
or to unset it, run:
  $ gcloud config unset project

Hmmm : so this seems to be untrue :

"Once connected to the cloud shell, you'll see that you are already authenticated 
and the project is set to your PROJECT_ID:"

Attempt to fix it (using the project id in the sidepane that the instructions told me to hide)

$ gcloud config set project qwiklabs-gcp-84f8f1c4f27f96a8
Updated property [core/project].

$ gcloud config list project
[core]
project = qwiklabs-gcp-84f8f1c4f27f96a8
Your active configuration is: [cloudshell-15818]

This looks better.

Enable Dataflow API

"On the Console, enable the Dataflow API by selecting APIs & Services > Library from the left menu"

So this does not refer to the Console terminal. But apparently the whole cloud webpage is the Console.

"On the dashboard search  for 'dataflow'."

Means type ‘dataflow’ into the search box. One panel called “DataFlow API - Google” comes up, mostly hidden by the Terminal Console.

Click on the overall panel : Something seems to happen (i.e. waiting signs). A “ENABLE” button appears. And a “Try this API” button too.

Click “ENABLE” per the instructions (though “trying” sounds appealing)

Now at “Library” page which says :

APIs & Services
Library
To view this page, select a project.

CREATE-button

This isn’t mentioned in the instructions, and doing “CREATE” seems to suggest the wrong project id. So don’t follow that route.

Launch Cloud Datalab (for real this time?)

$ gcloud compute zones list
NAME                       REGION                   STATUS  NEXT_MAINTENANCE  TURNDOWN_DATE
us-east1-b                 us-east1                 UP
us-east1-c                 us-east1                 UP
us-east1-d                 us-east1                 UP
us-east4-c                 us-east4                 UP
us-east4-b                 us-east4                 UP
us-east4-a                 us-east4                 UP
us-central1-c              us-central1              UP
us-central1-a              us-central1              UP
us-central1-f              us-central1              UP
us-central1-b              us-central1              UP
us-west1-b                 us-west1                 UP
us-west1-c                 us-west1                 UP
us-west1-a                 us-west1                 UP
europe-west4-a             europe-west4             UP
europe-west4-b             europe-west4             UP
europe-west4-c             europe-west4             UP
europe-west1-b             europe-west1             UP
europe-west1-d             europe-west1             UP
europe-west1-c             europe-west1             UP
europe-west3-b             europe-west3             UP
europe-west3-c             europe-west3             UP
europe-west3-a             europe-west3             UP
europe-west2-c             europe-west2             UP
europe-west2-b             europe-west2             UP
europe-west2-a             europe-west2             UP
asia-east1-b               asia-east1               UP
asia-east1-a               asia-east1               UP
asia-east1-c               asia-east1               UP
asia-southeast1-b          asia-southeast1          UP
asia-southeast1-a          asia-southeast1          UP
asia-southeast1-c          asia-southeast1          UP
asia-northeast1-b          asia-northeast1          UP
asia-northeast1-c          asia-northeast1          UP
asia-northeast1-a          asia-northeast1          UP
asia-south1-c              asia-south1              UP
asia-south1-b              asia-south1              UP
asia-south1-a              asia-south1              UP
australia-southeast1-b     australia-southeast1     UP
australia-southeast1-c     australia-southeast1     UP
australia-southeast1-a     australia-southeast1     UP
southamerica-east1-b       southamerica-east1       UP
southamerica-east1-c       southamerica-east1       UP
southamerica-east1-a       southamerica-east1       UP
europe-north1-a            europe-north1            UP
europe-north1-b            europe-north1            UP
europe-north1-c            europe-north1            UP
northamerica-northeast1-a  northamerica-northeast1  UP
northamerica-northeast1-b  northamerica-northeast1  UP
northamerica-northeast1-c  northamerica-northeast1  UP

Wowza, that’s a long list that scrolls off the page quickly.

How do I pick a zone? I need to open a zone/capabilities list to cross-reference, let’s pick the Asia Pacific Region to have a look.

But I need to know whether I’m going to be doing :

  • Training (DEFINITELY)
  • Training with GPUs (SOUNDS NICE)
  • Online Prediction (HMMM - aren’t we going to be predicting Baby Weights Online?)
  • Batch Prediction

Nothing say “ML Engine” - but nor do the other regions, so it’s not just me.

Let’s try the first one that looks good “asia-east1” (for ‘Training’, but it’s not ticked for ‘Online Prediction’, which is worrying).

$ datalab create babyweight --zone asia-east1
Creating the network datalab-network
#  Awkward pause 
Creating the firewall rule datalab-network-allow-ssh
#  Awkward pause 
Creating the disk babyweight-pd
ERROR: (gcloud.compute.disks.create) Could not fetch resource:
 - Invalid value for field 'zone': 'asia-east1'. Unknown zone.
A nested call to gcloud failed, use --verbosity=debug for more info.

Ahhh - perhaps I should look up something in the first column that matches the Region given as a column heading in the Available Regions table: ‘asia-east1-b’ fits the bill.

$ datalab create babyweight --zone asia-east1
Creating the disk babyweight-pd
Creating the repository datalab-notebooks
Creating the instance babyweight
Created [https://www.googleapis.com/compute/v1/projects/qwiklabs-gcp-84f8f1c4f27f96a8/zones/asia-east1-b/instances/babyweight].
Connecting to babyweight.
This will create an SSH tunnel and may prompt you to create an rsa key pair. To manage these keys, see https://cloud.google.com/compute/docs/instances/adding
-removing-ssh-keys
Waiting for Datalab to be reachable at http://localhost:8081/
This tool needs to create the directory
[/home/google646468_student/.ssh] before being able to generate SSH keys.
Do you want to continue (Y/n)?

This looks a lot more promising.

“You will get a SSH warning. Click Y to continue, and Enter through the passphrase questions. Datalab is ready when you see a message prompting you to use Web Preview to start using Datalab.”

Generating public/private rsa key pair.
Enter passphrase (empty for no passphrase):
Enter same passphrase again:
Your identification has been saved in /home/google646468_student/.ssh/google_compute_engine.
Your public key has been saved in /home/google646468_student/.ssh/google_compute_engine.pub.
The key fingerprint is:
SHA256:d10RTDpGREALLY?JHma1QEdMi2QFRCJj1rQ google646468_student@cs-6000-devshell-vm-5c0nfeffe-b2bd-43b4-a8cc-83d3fa4fc485
The key's randomart image is:
+---[RSA 2048]----+
|       =ooo*==++.|
|      o o.o +oo+.|
|         E o.*+ .|
|          .0O0.. |
|        S .+++.  |
|         . oo .  |
|          .+B.   |
|          o%o+   |
|          =+*o.  |
+----[SHA256]-----+
Updating project ssh metadata.../Updated [https://www.googleapis.com/compute/v1/projects/qwiklabs-gcp-84f8f1c4f27f96a8].
Updating project ssh metadata...done.
Waiting for SSH key to propagate.
#  Awkward pause 

#  Eventually this appears (while trying to do the next steps):
The connection to Datalab is now open and will remain until this command is killed.
Click on the *Web Preview* (square button at top-right), select *Change port > Port 8081*, and start using Datalab.

“Datalab will take about 5 minutes to start.”

Ok. so let’s keep going

“In the Console, go to Products & services > Storage. Create a bucket that is regional, and match the Location to region as the VM you just made. “

Hmmm. Doesn’t seem to be anything called that in the Left Hand Side panel.

How about “Select a Project” from the top bar? There are two projects available :

  • QwikLab resources
  • A project id that could be mine (Clicked this one)

Not sure that did anything. Nothing has changed on the Left Sidebar.

Next attempt : “Storage-Browser”

  • Leads me to “Buckets” with a “Create Bucket” button. Looks promising
  • “Create a bucket” : “Name Must be unique across Cloud Storage. If you’re serving website content, enter the website domain as the name.”
  • Flying blind now - pick a bucket name at random “random-bucket-name-please”
  • Regional (Checked) and “asia-east1” picked from long selection box
  • “Create”-button (Clicked) : Some activity
  • Now at a kind of upload/file list page

“Use the + button to open another tab of Cloud Shell” : Ahh - why didn’t you tell me about this before?

Code template to enter the bucket name wraps…

# gsutil cp gs://cloud-training-demos/babyweight/preproc/* gs://<BUCKET>/babyweight/preproc/
$ gsutil cp gs://cloud-training-demos/babyweight/preproc/* gs://random-bucket-name-please/babyweight/preproc/

Copying gs://cloud-training-demos/babyweight/preproc/eval.csv-00000-of-00012 [Content-Type=text/plain]...
Copying gs://cloud-training-demos/babyweight/preproc/eval.csv-00001-of-00012 [Content-Type=text/plain]...
Copying gs://cloud-training-demos/babyweight/preproc/eval.csv-00002-of-00012 [Content-Type=text/plain]...
Copying gs://cloud-training-demos/babyweight/preproc/eval.csv-00003-of-00012 [Content-Type=text/plain]...
/ [4 files][154.1 MiB/154.1 MiB]    1.0 MiB/s
==> NOTE: You are performing a sequence of gsutil operations that may
run significantly faster if you instead use gsutil -m -o ... Please
see the -m section under "gsutil help options" for further information
about when gsutil -m can be advantageous.
Copying gs://cloud-training-demos/babyweight/preproc/eval.csv-00004-of-00012 [Content-Type=text/plain]...
/ [4 files][154.1 MiB/321.6 MiB]    1.0 MiB/s

# VERY LONG WAIT

\ [55 files][  5.7 GiB/  5.7 GiB]    5.5 MiB/s
==> NOTE: You are performing a sequence of gsutil operations that may
run significantly faster if you instead use gsutil -m -o ... Please
see the -m section under "gsutil help options" for further information
about when gsutil -m can be advantageous.
Operation completed over 55 objects/5.7 GiB.

That looks like a warning that I’m doing something wrong - but the instructions don’t mention anything about this step. So I’ll just let it continue

Just a mo : There are only 23 minutes left to complete this thing…

Copying these files is like several Gb of data. And the speed sucks. Is it normally this slow?

Why are there 43 files of babyweight data? I just want to play around with a DataLab.
And this QwikLab is so slow & confusing that I’ll run out of time before even getting to the notebook.

Checkout notebook into Cloud Datalab

Will something happen now? Only … minutes left : Except the counter has reset to 1:30:00 (though there should be about 15mins left) - probably because I backbuttoned, and the state was lost.

“Web Preview > Change port.” - to 8081 : (DONE)

Ahah! The DataLab thing is working!

“In Datalab, click on the icon for Open ungit in the top-right.” : UnGit??? (DONE)

“remove “/notebooks” from the title of the page,” (DONE)

‘Clone From’ (https://github.com/GoogleCloudPlatform/training-data-analyst) (DONE)

Blimey : Stuff has been going on in this repo. (Best not to touch)

Back in the DataLab file page : “Jump to File” : “training-data-analyst/blogs/babyweight/train_deploy.ipynb” (DONE)

Execute training and prediction jobs

Set up the Bucket, Project and Region ids.

BUCKET = 'random-bucket-name-please'
PROJECT = 'qwiklabs-gcp-84f8f1c4f27f96a8'
REGION = 'asia-east1'

But doesn’t the cell “gsutil -m cp -R gs://cloud-training-demos/babyweight gs://${BUCKET}” essentially re-do what I did on the command line (which took ages)?

# ... long wait ...
Copying gs://cloud-training-demos/babyweight/trained_model_tuned/model.ckpt-571432.index...
Copying gs://cloud-training-demos/babyweight/trained_model_tuned/model.ckpt-571432.meta...
- [609/609 files][  6.1 GiB/  6.1 GiB] 100% Done  12.6 MiB/s ETA 00:00:00       
Operation completed over 609 objects/6.1 GiB.    

How am I meant to get through this notebook in 15 minutes?

Carry out local training

Just pressing Ctrl-Enter repeatedly now, since there’s no time to read the cells…

import tensorflow cell :

/usr/local/envs/py2env/lib/python2.7/site-packages/h5py/__init__.py:36: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from ._conv import register_converters as _register_converters
/usr/local/envs/py2env/lib/python2.7/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
/usr/local/envs/py2env/lib/python2.7/site-packages/h5py/_hl/group.py:22: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .. import h5g, h5i, h5o, h5r, h5t, h5l, h5p
/usr/local/envs/py2env/lib/python2.7/site-packages/scipy/ndimage/measurements.py:36: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from . import _ni_label
/usr/local/envs/py2env/lib/python2.7/site-packages/simplejson/encoder.py:286: DeprecationWarning: Interpreting naive datetime as local 2018-07-04 16:30:35.570388. Please add timezone info to timestamps.
  chunks = self.iterencode(o, _one_shot=True)

At the training step :

/usr/local/envs/py2env/lib/python2.7/site-packages/sklearn/utils/__init__.py:10: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .murmurhash import murmurhash3_32
/usr/local/envs/py2env/lib/python2.7/site-packages/sklearn/utils/extmath.py:24: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from ._logistic_sigmoid import _log_logistic_sigmoid
/usr/local/envs/py2env/lib/python2.7/site-packages/sklearn/metrics/cluster/supervised.py:23: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .expected_mutual_info_fast import expected_mutual_information
/usr/local/envs/py2env/lib/python2.7/site-packages/sklearn/metrics/pairwise.py:30: RuntimeWarning: numpy.dtype size changed, may indicate binary incompatibility. Expected 96, got 88
  from .pairwise_fast import _chi2_kernel_fast, _sparse_manhattan

Looks like something happened :

!ls babyweight_trained
checkpoint				     model.ckpt-1000.index
eval					     model.ckpt-1000.meta
events.out.tfevents.1530721891.7286677d5bcc  model.ckpt-1.data-00000-of-00001
export					     model.ckpt-1.index
graph.pbtxt				     model.ckpt-1.meta
model.ckpt-1000.data-00000-of-00001

Carry out distributed training

Is that’s what’s happening here? :

Removing gs://random-bucket-name-please/babyweight/trained_model/model.ckpt-390628.meta#1530721709406785...
/ [58/58 objects] 100% Done                                                     
Operation completed over 58 objects.                                             
Job [babyweight_180704_163436] submitted successfully.
Your job is still active. You may view the status of your job with the command

  $ gcloud ml-engine jobs describe babyweight_180704_163436

or continue streaming the logs with the command

  $ gcloud ml-engine jobs stream-logs babyweight_180704_163436
/usr/local/envs/py2env/lib/python2.7/site-packages/simplejson/encoder.py:286: DeprecationWarning: Interpreting naive datetime as local 2018-07-04 16:34:36.878487. Please add timezone info to timestamps.
  chunks = self.iterencode(o, _one_shot=True)

… No idea, the connection just timed out.

“The connection to your Google Cloud Shell was lost. “

Deploy the ML model as a web service (OUT OF TIME)

Make predictions with the model (OUT OF TIME)


LEARNING FAILURE



Martin Andrews

{Finance, Software, AI} entrepreneur, living in Singapore with my family.



blog comments powered by Disqus