Panagiotis Braimakis
Panagiotis Braimakis
While playing with the parameters in code i noticed that the silence splitting works better for mono vs stereo wav files. Does this make sense ? Thanks pb
Would it be possible to do transfer learning to detect cars with this model ? How much data would be necessary to do so?
The title is self explanatory! Could you share also the notebook you have used to calibrate the fbank_net ? It seems (as per your notes it performs better) Thanks pb
The files seems to split randomly the dataset, a 25-75 split normally should have representatives from each category. This is not the case. If for instance a category exists in...