TolgaBkm

Results 5 comments of TolgaBkm

RuntimeError: DataLoader worker (pid(s) ****, ****, ****) exited unexpectedly

These steps solved my issue: 1. Set `cfg.data.workers_per_gpu=0` 2. Open mmseg>datasets>builder.py 3. Change Line 99 as `persistent_workers=False`

SSL error when trying to load ImageNet weights for se_resnext101_32x4d encoder with any backbone.

This solution worked for me: https://github.com/qubvel/segmentation_models.pytorch/issues/144#issuecomment-585754975

Dataset download question

You can fill the data request form here: http://www2.isprs.org/commissions/comm3/wg4/data-request-form2.html

EfficientDet Training loss=nan

I have the same problem with EfficientDet D4. ![D4](https://user-images.githubusercontent.com/57003251/91416865-a61e6f80-e858-11ea-8aa9-4a3076a19a1e.PNG)

EfficientDet Training loss=nan

I have also tried to increase the batch size to 4. However, I still got loss=NaN around 50k steps of training.