Deploying an English Speech to Text model to heroku

Abraham Jerry Kakooza — Wed, 16 Mar 2022 17:12:42 +0000

I will be showing you how to deploy the coqui-ai english speech to text model on heroku such that you can use it within your applications.

The model we will deploy can be found here . It is a large vocab English speech to text model.

First within your terminal, clone this repository as shown below and navigate into the en-stt-engine directory

git clone https://github.com/KakoozaJerry/en-stt-engine.git
cd en-stt-engine

Make sure you have git lfs installed on your machine, you can find instructions here.

Run the command below after installing git lfs

git lfs install

It should return something like what is below

Updated git hooks.
Git LFS initialized.

Open your heroku in the browser and type the link below

https://dashboard.heroku.com/new-app

Incase you are not logged in you will be prompted to do so.
Your webpage should return the view below, where you can input the name you want for the app, I use english-stt

Now you can go back to your terminal and run

heroku login

You will see the log below

heroku: Press any key to open up the browser to login or q to exit:

Press any key as stated and click the login button shown in the browser, if successful this will show.

Opening browser to https://cli-auth.heroku.com/auth/cli/browser/......
Logging in... done
Logged in as user@gmail.com

Set the name of the remote origin from heroku app you created above using the command below

heroku git:remote -a <stt-app-name>

This should return

set git remote heroku to https://git.heroku.com/<stt-app-name>.git

Specify the heroku python as your build pack as shown below in your terminal

heroku buildpacks:set heroku/python

This is what it should return

Buildpack set. Next release on english-stt will use heroku/python.
Run git push heroku main to create a new release using this buildpack.

Add the ffmpeg buildpack since this will also be needed

heroku buildpacks:add https://github.com/jonathanong/heroku-buildpack-ffmpeg-latest.git

Below is what it should return

Buildpack added. Next release on english-stt will use:
  1. heroku/python
  2. https://github.com/jonathanong/heroku-buildpack-ffmpeg-latest.git
Run git push heroku main to create a new release using these buildpacks.

Add git-lfs buildpack using the command below

heroku buildpacks:add https://github.com/raxod502/heroku-buildpack-git-lfs

This is what will be returned

Buildpack added. Next release on english-stt will use:
  1. heroku/python
  2. https://github.com/jonathanong/heroku-buildpack-ffmpeg-latest.git
  3. https://github.com/raxod502/heroku-buildpack-git-lfs
Run git push heroku main to create a new release using these buildpacks.

Set your HEROKU_BUILDPACK_GIT_LFS_REPO variable to point to the github repository as shown below

heroku config:set HEROKU_BUILDPACK_GIT_LFS_REPO=https://github.com/KakoozaJerry/en-stt-engine.git

This is what will be returned

Setting HEROKU_BUILDPACK_GIT_LFS_REPO and restarting ⬢ english-stt... done, v6
 ▸    Warning: The "HEROKU_" namespace is protected and shouldn't be used.

You should be able to publish your app using this command

git push heroku main:master --no-verify

This will take a while depending on the speed of your internet

After a successful deployment, you should see successful deployment and you can test out your stt api in postman as shown below

You can test it out with this audio and a key value of file.

Make sure to select form data as shown in the screenshot above.

DEV Community: Abraham Jerry Kakooza

Deploying an English Speech to Text model to heroku