Implement audio classifier #31
Conversation
|
I know this is a WIP, so keeping comments light. In the Library fragment, the UI should be cleared if backgrounding the app and returning. Can we float the overlap percentage to the right so it aligns with our inference time? Other than those incredibly small things, the UI looks solid to me. Did you have any luck with the streaming inference time values based on our earlier conversation? Thanks! |
Unfortunately, the timestamp |
ghost
changed the title
| } | ||
|
|
||
| when (currentModel) { | ||
| YAMNET_MODEL -> { |
There was a problem hiding this comment.
Since we're only using one model and not showing a radio button for changing it in our UI, let's remove this when statement and related constants. We can readd them later if we support an additional model.
.../app/src/main/java/com/google/mediapipe/examples/audioclassifier/fragment/LibraryFragment.kt
Outdated
Show resolved
Hide resolved
|
Very minor comments, otherwise LGTM. I'll see if I can get some answers about the inference time, but that may not happen until next week. I'll update here once I have some more information. Thanks! |
|
Would you mind moving this app one level lower? It'd go into an android directory to match the structure from the other samples. |
ghost
requested a review
from …io_classifier Implement audio classifier
No description provided.