MemoQ + AutoML integration

[Sebastián Martín]

Hi Community!

I've been struggling to connect MemoQ and AutoML. I've trainer several engines in AutoML, created one service account with the following roles:

- AutoML Service Agent

- Cloud Translation API Admin

- Cloud Translation API Service Agent

- Cloud Translation API User

I downloaded the JSON file from the KEY section of the service account but when I upload the JSON within memoq I receive the following message:

"No custom model available"
"Google will use the general model"

Any thoughts?

Thanks in andvance!

Sebastián

[Adam Bittlingmayer]

You should try to understand if it's an issue with the Google model and API key or just a memoQ issue.

(Note that using a custom Google Translate model requires 2 pieces of info: the *model path* and the API key JSON.  The JSON alone isn't enough.)

To check that the Google model and auth is working, you can try one of the example code snippets.  e.g. https://cloud.google.com/translate/automl/docs/samples/automl-translate-predict#automl_translate_predict-python (But it is also a pain to get your key in your environment.)

Or if you don't want to code, you can go to console.modelfront.com and create a new free account and an evaluation, and paste in your Google model and API key JSON.

[Sebastián Martín]

Hi Adam,

Thanks for your input. I'll try to create a modelfront test.

Can you help me to complete this?

projects/<project-number-or-id>/locations/<zone>/models/<model>

I have the project name and the name of the trained model but where, in the google cloud console,  could I get the specific zone for this engine?

Thanks!

Sebastián.

[Adam Bittlingmayer]

Go to https://console.cloud.google.com/translation/models/list

Then select the project and model, and scroll down to the section "Use your AutoML model".

Of course, you must do this per model i.e. per language.