Monthly Archives: April 2019

IBM Watson Services : what is included in your free account ?

You can have a free account in IBM Cloud without time limitations and only usage limitation. Let’s see what you can do with this account and you will realize you can already do a lot with it !

One limitation to be aware is that a service is deleted after 30 days of inactivity. You will get an email to warn you of the deletion a few days before. If you want to keep using it, just go to the service and use it once for another 30 days.

I will list what is included in each Watson services to show you you can start right now to develop with IBM before even starting to pay for anything.

Watson Assistant (WA)

Watson Assistant helps you develop dialogs and chatbot for any kind of use case. Here is what you can do.

  • 10,000 Messages/Month
  • AI-Based Intent and Entity Recognition
  • Entity Synonym Recommendations
  • Visual Dialog Edit with Simple Response Types (Text, Options, Images, etc…)
  • Prebuilt Content Available
  • Analytics Dashboard with 7 Days of Storage
  • 5 Dialog Skills, Each with 100 Dialog Nodes
  • Shared Public Cloud

This is already a lot to start a dialog bot, test it and use it with your firsts users.

Main part missing in the Lite plans :

  • Discovery skills that allows to go to Discovery for search directly from Assistant (Short Tail / Long Tail made easy)
  • Disambiguation
  • Intent Recommendations
  • Intent Conflict Detection
  • Service Desk Integrations

Watson Discovery (WDS)

Watson Discovery helps you index documents in a smart way and find useful information in them you couldn’t find another way. Here is what you can do.

  • 1,000 documents per month
  • 1 custom model
  • 500 element classification pages per month

A custom model is a new context model developed with Watson Knowledge Studio.

Visual Recognition

  • 1,000 Events per month towards:
    • Pre-trained model classification (General, Face, Food, Explicit) (images)
    • Custom Model classification (images)
    • Custom Model training (images)
    • 2 Custom Models
    • 1 Lite Plan instance per IBM Cloud Organization
    • Free Exports to Core ML

Language Translator (LT)

Language Translator helps you translate language in another language

  • 1,000,000 Characters per Month

Natural Language Understanding (NLU)

  • 30,000 NLU Items Per Month
  • 1 Custom Model

NOTE: A NLU item is based on the number of data units enriched and the number of enrichment features applied. A data unit is 10,000 characters or less. For example: extracting Entities and Sentiment from 15,000 characters of text is (2 Data Units * 2 Enrichment Features) = 4 NLU Items. A custom model refers to an annotation model developed with Watson Knowledge Studio.

Personality Insights (PI)

  • 1,000 API Calls per Month

Speech To Text (STT)

  • 100 Minutes per Month

Text To Speech (TTS)

  • 10,000 Characters per Month

Tone Analyzer

  • 2,500 API Calls per Month

How to start with IBM Cloud and IBM Watson : Create your account.

You may have seen some nice presentation on IBM Cloud and IBM Watson and want to try now yourself ? It’s very easy and you can get a lot of services for free to start your projects.

IBM Cloud registration

IBM Cloud registration is easy and give you access to 61 services available in lite mode and covers the full range of services : Development, Storage, Databases, AI with Watson, API, …

Here is how to register. It is very simple. Go to and click on “Create an IBM Cloud account”. Fill in your email address and all details asked and click on “Create Account”. You may have a Captcha to answer, right now we don’t allow AI to create their own account 🙂

You will get an email to confirm your email address.

Click on “Confirm account”and here we go.

Click on “Log In”

You then have to accept the IBMid Account Privacy policy by clicking on “Proceed” after reading them.

You have now access to the IBM Cloud Dashboard !

Let’s see in the next post, what you can use for free and all the things you can play with.

How to upgrade TensorFlow in Watson Studio Jupyter Notebook

I recently wanted to learn more about TensorFlow and as a Watson Solution Architect, I am using Watson Studio Jupyter Notebooks. On my first test I had an error on the second command.

I realized I was using an old version. Let’s see how to upgrade to a newer version. Watson Machine Learning is using TensorFlow version 1.11 so I wanted to use the same one. Here are the steps.

#Install a pip package in the current Jupyter kernel
import sys
  !{sys.executable} -m pip install tensorflow==1.11.0

This will install tensorflow version 1.11. It should whow the download of the version. To check the version number, here is the command.

import tensorflow as tf

If you don’t get 1.11.0, you may need to restart the kernel.

Now everything is working.

Next step : configure directly an environment to select the components version.