Adding additional models with argument

[jacksongoode]

I see within inference.py...

 # Model
        if model is None:
            self.model = Cnn14(sample_rate=32000, window_size=1024, 
                hop_size=320, mel_bins=64, fmin=50, fmax=14000, 
                classes_num=self.classes_num)
        else:
            self.model = model

But I was wondering if you could include the possibility of adding other models to your model.py? I am attempting to use the lightweight MobileNetV2.

[jacksongoode]

I managed to copy the required functions from the models.py defined in your audio_taggin_ccn repo, but I am not getting any output from the example described in the readme of this repo.

(clipwise_output, embedding) = at.inference(audio)

Does this print anything on its own? Or does it need to be called within print?

[jacksongoode]

I was able to get it working checking out your example.py. Really wonderful stuff and these lightweight models are particularly useful in real-time environments. I would include these functions (in example.py) within the pip package so that it's a little more intuitive to run as a function simply.

[qiuqiangkong]

Thanks! We will consider to put the MobileNet to the repo.