-
![截屏2023-06-18 19 25 57](https://github.com/dingmyu/davit/assets/69461641/777cef2f-c428-44a0-a9ec-0cdbf9ade8c5)
When I used the mmdet framework, I set the parameter model.pretrained to a pretrained m…
-
-
### Bug description
Lightning throws an error when using a saved model to run inference, while using the ddp_notebook strategy.
In this case, it throws the error: "RuntimeError: Lightning can't cr…
-
biigle/volumes#55 will implement support for standardized image metadata fields. Among these is `REF_laserdistances` with the laser distance (in m? ask Timm). If this field is available for an image, …
-
I am using ubuntu 22.04, I had many errors and spent a whole day getting this to work. I will outline the steps I needed to take to get it working. hopefully this helps if you get any of the errors me…
-
Thank you for sharing the code for your great work!
I had this problem when I tried to run demo_imgs.py!
self.act1 = model.act1
AttributeError: 'EfficientNetFeatures' object has no attribute 'act1'…
-
Hello, could you please provide ViT-B-16's weights pretrained on CIFAR-100 dataset? Many thanks!
-
Hi,
I understand the visformer accept input in the 3D format as in (3,244,244), can I feed data with more than 3 RGB channel into the model? like in (16, 244,244)?
-
-
```
gtail on ocikbgtw:
Does not show (only) the stderr of the job, if -e option is applied,but (also)
the stdout.
I compared both gtail-outputs, one with -e and one with -o option:
No difference fo…