Robin Cole
Robin Cole
Similarly for the crop classification using the [supplied config](https://github.com/NASA-IMPACT/hls-foundation-os/blob/main/configs/multi_temporal_crop_classification.py), I've achieved slightly lower metrics than those reported, although I did have to reduce the batch size. ``` 2024-01-02 16:56:40,546 -...
@NielsRogge why close this? `As can be seen, the per_category_iou returns nan for categories which aren't present in the preds and targets, taking into account the ignore_index.` This is essential...
`save=True` would be a good default, that appears to be the behaviour when using CLI & `yolo task=detect mode=predict`
Quite surprised these are not default `true`
It is possible to embed the YouTube in the readme, if of interest
See end of https://github.com/robmarkcole/HASS-amazon-rekognition
It will display the video thumbnail,which can be whatever you like
Another option is to make a completely new video, just a quick intro to torchgeo, and use that
It is pretty straightforward in 2 steps: (1) process the dataset into the required binary format and host on a cloud - I am using AWS, TBC which could would...
I'm not sure there is a way around it. In my implementation the data module accepts a 'streaming' arg and returns the regular or streaming dataset