valentin7121 issues

Results 5 issues of


                                            valentin7121

Feature request: annotation transparency

Depending on the brightness and contrast of the annotated images, different opacity of bounding box is optimal and comfortable. Can such a setting be added for categories?

Fix endianness

Set platform independent endianness in struct.pack()

Pre-trained model link

Could you please upload pre-trained model on Google Drive?

The article ["Conformer-based Target-Speaker Automatic Speech Recognition for Single-Channel Audio"](https://arxiv.org/pdf/2308.05218v1.pdf) states that the model CONF-TSASR will be open-sourced through NVIDIA NeMo toolkit. Is it already open-sourced? Will pretrained weights be...

R_n in rake_mvdr_filters with real data

In [example](https://github.com/LCAV/pyroomacoustics/blob/master/examples/beamforming_time_domain.py#L96C5-L96C40) `R_n` in `rake_mvdr_filters` is set to `R_n=sigma2_n * np.eye(mics.Lg * mics.M)`. I guess that `R_n` is the noise covariance matrix and when processing a signal from real microphones...