ncnn android with 1 channel y model

[mudasarK]

Hello,
I am experimenting with ncnn in my android app to replace tensorflow lite.
My models most of time use Y channel (1C) instead of RGB (3channel), reason for it is speed.

When I am running my 1 channel models on my pc application code (QT application) using ncnn, results are fine.
But when I run that model on android 11 it show broken pixels. (please refer to image below).

In android, I have a shader code which converts Bitmap from BGRA (little endia) to Y channel and which I am feeding to ncnn model like this

void *data = (void *)env->GetDirectBufferAddress(image_buffer);
ncnn::Mat in = ncnn::Mat(w, h, c, data, (size_t)4u);
.
.
ncnn::Mat out;
ret = ex.extract(oidxs[0], out);

and then I am returning results like this
jobject obj = env->NewDirectByteBuffer(out.data, (out.wout.hout.c*out.elemsize));

and after results are return my code is converting Y back to BGRA to draw on app.

you can see whole white background is looking like a net with holes.

Note: I am using same shader and same model with tensor-flow lite on same application and results are fine.
am I missing something?

P.s. I tried 3 channel model also when I feed data using
ncnn::Mat in = ncnn::Mat(w, h, c, data, (size_t)4u);
instead of
ncnn::Mat::from_android_bitmap
it shows same results as above.

Please guide

Thank you in advance

[mudasarK]

This fixed the issue
opt.use_fp16_packed = false;
opt.use_fp16_storage = false;
opt.use_fp16_arithmetic = false;