Latest activity

Perfect. Yes, there are a couple of changes in Android 13 preventing the lib from working correctly on a number of Dimensity-based...

I ran accel. int8 test on mobilenetv2, and it worked well. I'm going to run the entire test. So, is it because Neuron lib is obsolete...

Hi @jinqimu, Thanks for the provided info. Can you please try to install and run this beta benchmark build with an updated Neuron lib...

phone model: vivo IQOO Neo7 SoC: mediatek diversity 9000 plus Android version: 13 kernel version: android12-9

Detailed AI Benchmark V5 results were released for over 750 Android devices: https://ai-benchmark.com/ranking...

That's interesting, can you please specify your phone model and Android version/build? An info from the logcat for the 1st test would...

Try to set this option to true.

So i wanted to benchmark my own custom model on a android 9, 675 phone using just the NNAPI, so i used this app to benchmark the model...

My phone's soc is mediatek diversity 9000+. Mobilenetv2 with mediatek neuron delegate doesn't work on my phone. For this reason, the...

* TF Version: 2.10.0 * Platform: Windows-10-10.0.19045-SP0 * CPU: N/A * CPU RAM: 32 GB * GPU/0: NVIDIA GeForce RTX 4090 * GPU RAM...

* TF Version: 2.10.0 * Platform: Windows-10-10.0.19045-SP0 * CPU: N/A * CPU RAM: 32 GB * GPU/0: NVIDIA GeForce RTX 4090 * GPU RAM...

Thank you very much.

Thanks for the reply. This benchmark is running in Ubuntu, under WSL2. As far as I understand, it should be using CUDA directly, not...

Yes, that's normal - these models are using specific NLP-based ops that are currently not fully supported by DirectML.

5667 GPUV