Emily (Miller) Dorne
Emily (Miller) Dorne
We should be able to specify the learning rate the same way Lightning does after it finds the desired one with [auto_lr_find](https://pytorch-lightning.readthedocs.io/en/stable/advanced/training_tricks.html#learning-rate-finder). It's simpler to do it right before training...
The README and error raised say that the sensitive attribute column should be in categorical, but passing in a categorical type fails because [the code is actually checking if the...
To reproduce, set up a conda environment with python=3.12 and then try to pip install the package. Unpinning numpy should resolve this. Is there a reason this specific version is...
Currently we support 3.8 and 3.9 but these are nearing [end of life](https://devguide.python.org/versions/). We should update docs and tests to use 3.10, 3.11, and 3.12.
As of version 1.9, when Lightning loads a checkpoint, it now applies a version migration on-the-fly, but it does not modify your checkpoint file. First, we should determine if this...
Running `load_video_frames` on different operating systems leads slightly different values. For example, vs. on linux vs. mac with the same ffmpeg version (4.4.3) and same python environment. Value differences are...
For distance estimation, if there are multiple detections in the frame, the distance estimate is repeated (but the same for both). This is because the input used for each is...
I'm geocoding a country level dataset on maternal mortality indicators. All countries worked fine except for Mexico which seems to map to a polygon only around the capital and not...
Goal: support users who want estimates across an entire body of water (e.g. heatmap-like output) This would entail a preprocessing step where we: - accept a polygon as input -...
We should let users specify a cache directory for downloading the imagery. This entail adding a `cache-dir` CLI argument and passing this to [`CyFiPipeline`](https://github.com/drivendataorg/cyfi/blob/1c9a61a965618379a281b37a08b9d0611a08280d/cyfi/pipeline.py#L24-L31) in [`predict`](https://github.com/drivendataorg/cyfi/blob/1c9a61a965618379a281b37a08b9d0611a08280d/cyfi/cli.py#L119) and [`predict_point`](https://github.com/drivendataorg/cyfi/blob/1c9a61a965618379a281b37a08b9d0611a08280d/cyfi/cli.py#L165) and then...