Benjamin Weinstein

I'm waiting on this as well. All downstream packages that require this repo cannot go on conda. https://github.com/conda/conda-build/issues/548

Normally, you use the class object of the deepforest model for evaluate. evaluate.evaluate is more of an internal function. How can we improve docs to show this? If you look...

I don't have any reason to believe that geotif would be different than .png, can you share the geotiff version? @easz is right, it all depends on patch size, which...

Just as a sanity check, can you share your annotations file, I might have a malformed box or class. I'd be interested in exploring this more, retinanet should be formatting...

@dassaniansh let's keep the convo here. See the link above. The use_bird_release model has 1 class 'Bird', imagine a user wants to start from that model, but has 5 classes...

No model training is needed. To create a deepforest model, there is a classes argument. https://github.com/weecology/DeepForest/blob/02aac1fd8105f18e63c321be170c2c55d82a51c6/deepforest/main.py#L23 so we know at time of model creation how many classes are desired. But...

Looks good, you just need to add the 'image_path' as one of the asserted columns to make sure all tests pass. Here: https://github.com/dassaniansh/DeepForest/blob/872364a40d392d1d67bd46b3cbcfbe2bb713c078/tests/test_main.py#L129 https://github.com/dassaniansh/DeepForest/blob/872364a40d392d1d67bd46b3cbcfbe2bb713c078/tests/test_main.py#L142 I can then merge. Thanks for...

Added to #353

I need to document this better, I see the confusion. You need to define a transform for both train and eval in the same function. So its ``` def get_transform(augment):...

great. If you can, I strongly encourage you to link to issue #350 and add your dataset for others to benefit from in the future. The backbone can only get...