You will learn how to prepare a GitHub repository to be used in a CI/CD development flow.

Before you begin

You will need a GitHub account.

GitHub requires that a Personal Access Token be set. If you do not have this on your account already, navigate to Settings > Developer Settings > Personal Access Tokens, click on Generate new token, and save the token locally.

Ensure you have enabled the token to Update GitHub Action workflows.

Setup your AVH instance, following the instructions in the Arm Virtual Hardware install guide . An AWS account is required.

Fork the example repository

As you shall be making modifications to the reference example, you must make your own copy (fork) of the repository.

In a web browser, navigate to the repository at:

    

        
        
            https://github.com/ARM-software/AVH-TFLmicrospeech/fork
        
    

and create a fork in your own personal repository store (you must be logged into GitHub).

It is assumed below that you have used the same repository name (AVH-TFLmicrospeech) for your fork.

Clone this repository

In your Arm Virtual Hardware terminal, clone the fork of the repository, and navigate into its directory.

    

        
        
            git clone https://github.com/<Your_GitHub_Username>/AVH-TFLmicrospeech
cd AVH-TFLmicrospeech/Platform_FVP_Corstone_SSE-300_Ethos-U55
        
    

Build the example (optional)

If you wish to verify the project behaves as expected, rebuild the project:

    

        
        
            cbuild.sh --pack microspeech.Example.Reference.cprj
        
    

To run on the Virtual Hardware, use the provided script:

    

        
        
            ./run_example.sh
        
    

Observe the output:

    

        
        Heard yes (146) @1000ms
Heard no (145) @5600ms
...

        
    

Modify the example

Edit command_responder.cc source file, which defines the output message.

    

        
        
            nano ../micro_speech/src/command_responder.cc
        
    

For example, change Heard to The word was in the TF_LITE_REPORT_ERROR() function.

    

        
        
                TF_LITE_REPORT_ERROR(error_reporter, "The word was %s (%d) @%dms", found_command, score, current_time);
        
    

Rebuild and rerun

Rebuild and rerun the example.

    

        
        
            cbuild.sh microspeech.Example.cprj
./run_example.sh
        
    

Observe that the output has changed as expected.

    

        
        The word was yes (146) @1000ms
The word was no (145) @5600ms
...

        
    

Update the GitHub repository

Set your GitHub login details in your Virtual Hardware instance

    

        
        
            git config --global user.name "<GitHub_Username>"
git config --global user.email <Email>
        
    

Verify that the repository fork is referenced:

    

        
        
            git remote -v
        
    

Commit and push changed file(s) to the repository:

    

        
        
            git add ../micro_speech/src/command_responder.cc
git commit -m "changed output message"
git push
        
    

You will be prompted for your GitHub username and Personal Access Token (password).

Refresh your browser and observe that your fork of the repository has been updated appropriately.

Back
Next