Jake Graving

I implemented the changes in the `Annotator`. Before I push to pip could you check that it's working for you by installing the latest dev version? ``` bash pip install...

Good to hear! Currently scenario (1) is used at training time as the network has to be optimized to predict zeros when the keypoint isn't visible so the low confidence...

> Something that confused me at first, so maybe helps others to understand, is that the data generator converts NaN (non-visible) coordinates to -9999, presumably because this is so far...

>There are three places changes could be implemented: (1) Logger callback on_epoch_end; (2) base_model.evaluate; (3) utils.keypoints.keypoint_errors function. (3) makes the most sense to me, but wonder what you think? Also,...

This is definitely possible but we would want to avoid adding too much complexity to the code. The easiest approach is probably to add an option for the `TrainingGenerator` that...

I ran some tests and it looks like this is probably not worth implementing. The opencv resize function appears to be significantly faster on all counts. There's just a ton...

Shouldn't be too difficult to add, but it's not high priority at the moment. I'll need to think about how best to accomplish this. If you want to submit a...

Using opencv to resize images doesn't require any interaction with the `BaseModel` or the maxima layers. I would just modify the [`BaseGenerator`](https://github.com/jgraving/DeepPoseKit/blob/master/deepposekit/io/BaseGenerator.py#L125) with a `resize` kwarg (with `resize=None` as a...

Hi, Thanks for your interest! However, I'm a bit confused by your feature request, so I'd appreciate if you could provide more details. During model training and when annotating images,...

We've now added a (experimental) data generator for loading DeepLabCut data (see here for example: https://github.com/jgraving/DeepPoseKit/blob/master/examples/deeplacut_data_example.ipynb). The `BaseGenerator` class is now abstracted so writing a custom generator is possible with...