ctwillson

UP

Hi andrewmabc: Have you solved this problem?

> Not currently, but it shouldn't be too hard to add support for it. With the signal/ ops you can implement a mel-spectrum filter bank followed by a log. To...

> You would have to do the feature extraction within your pre-processing code, you can see an example of how to do this here: https://github.com/ARM-software/ML-examples/tree/506c941bebdeb55aedc1b8cc53f27c482cf67ec8/tflu-kws-cortex-m/kws_cortex_m/Source/MFCC or here: https://review.mlplatform.org/plugins/gitiles/ml/ethos-u/ml-embedded-evaluation-kit/+/refs/heads/main/source/application/api/use_case/kws/src/KwsProcessing.cc Thanks for...

> The ML-Zoo repository is only for providing ML models for people to use. It isn't focused on showing complete end to end embedded applications. > > The links I...

> @bdytx5 please pull the lates version of kws_streaming: I added conversion of model (with internal state streaming stored in saved_model) to tflite. > > There is no need to...

> > In [kws_streaming/layers](https://github.com/google-research/google-research/tree/master/kws_streaming/layers) including [svdf](https://github.com/google-research/google-research/blob/master/kws_streaming/layers/svdf.py) I avoid using any custom TFLite operations including fused [TFLite SVDF](https://www.tensorflow.org/lite/guide/ops_compatibility#experimental_operations). There are several reasons for that: 1 Benchmarks of [svdf](https://github.com/google-research/google-research/blob/master/kws_streaming/layers/svdf.py) was not that...

@kismeter Maybe u need to implement the "representative_dataset" function,and then u can set inference_input_type change to tf.int8. BTW,can u provide the requirements.txt.I cannot run the lastest kws code on my...

> > @kismeter Maybe u need to implement the "representative_dataset" function,and then u can set inference_input_type change to tf.int8. BTW,can u provide the requirements.txt.I cannot run the lastest kws code...