Can't build the project

[AndroidDeveloperLB]

IDE says I need to update gradle, so I have, and now I got this:

* What went wrong:
Failed to notify dependency resolution listener.
'void org.gradle.api.artifacts.DependencySubstitutions$Substitution.with(org.gradle.api.artifacts.component.ComponentSelector)'

What should I do?

ESRGAN-Android-TFLite-Demo-master.zip

[tumuyan]

1. this repo modify from https://github.com/tensorflow/examples/tree/master/lite/examples/super_resolution/android
   it is some years ago, I think it need a early gradle not the latest.
2. the repo only share the compiled apk to test the tflite model. The answer is that the model is much worse than real-esrgan.
3. if you want to study how to use tflite, you should visit https://github.com/tensorflow/example , if you need a super resolution App, you could vist https://github.com/tumuyan/RealSR-NCNN-Android

[AndroidDeveloperLB]

1. Probably. Sadly the model there is tiny and I think it was very restricted. I still don't get how to create anything related to Tensorflow-lite, and I tried many times and sadly there isn't much documentation about it with a sample/video of how to do it. Not only that, but I've read that Tensorflow-lite was replaced recently with something else, which has even less docs...
2. Do you know perhaps how to use a model of any kind for super-resolution? Is there any library that does it, that I give it a Bitmap and target resolution, and it produces a new Bitmap of it?
3. I tried this now (here), and it always fails to do anything with images:

It keeps saying the file doesn't exist, or something similar.

On the emulator, I tried old version of Android (API 27), and it accepted after many tries to reach a file on the "Pictures" (tried when the file is in other folders) folder, but even then still failed:

1.zip

I also can't find where do I put the scale or target image resolution that I want to have.

I think it needs to be updated too. Still, not sure how it works. Which files are needed, how to make it super minimal in code, and have the flexibility of which models to use, in case a new model out there seems to be better in the future.

Reported here:

tumuyan/RealSR-NCNN-Android#73

[tumuyan]

1. I double-checked the files on my computer. Unfortunately, I had deleted the local files after uploading the APK, so I am unable to provide information about the tools I used.
2. Some smartphone manufacturers provide SDKs for SR. For instance, Huawei offers an SDK for image super-resolution https://developer.huawei.com/consumer/en/hiai/engine/image-super-resolution . If you cannot understand what this repository does, it can be challenging to explain how to use a model.
3. I will reply you in another issue