chore: Add basic statistics to demo apps

[msluszniak]

Description

Added inference times, plus for some apps TTFT, tokens/second.
Fixed T2S streaming for unfinished sentences.

Introduces a breaking change?

• Yes
• No

Type of change

• Bug fix (change which fixes an issue)
• New feature (change which adds functionality)
• Documentation update (improves or adds clarity to existing documentation)
• Other (chores, tests, code style improvements etc.)

Tested on

• iOS
• Android

Testing instructions

• Run all apps and check if basic statistics are correctly displayed
• Check that text without end of sequence sign is processed correctly in text to speech example

Screenshots

Related issues

Closes #959

Checklist

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have updated the documentation accordingly
• My changes generate no new warnings

Additional notes

[barhanc]

Instance segmentation app is missing the inference timing.

[msluszniak]

Instance segmentation app is missing the inference timing.

And model listing as well. It's because instance segmentation was merged after the PR with demo app fixes. I'll add these to instance segmentation. Good catch.

DONE

[barhanc]

Additionally there are no inference timings for other speech models, but I understand it would be difficult to properly measure the inference time there as they operate in streaming mode. Maybe @IgorSwat has some idea how this could be done.

[msluszniak]

@barhanc about streaming tasks, yeah I eventually decided that I don't have good idea how to present these times

[IgorSwat]

Additionally there are no inference timings for other speech models, but I understand it would be difficult to properly measure the inference time there as they operate in streaming mode. Maybe @IgorSwat has some idea how this could be done.

We could measure the average time of executing transcribe() inside the streaming loop - the question is whether we want to make a breaking change just for that.

Same goes for Text to Speech - measuring it's inference time is possible, but would require digging into TypeScript API at very least.