Tobias Czempiel

extract_optical_flow using docker image

9

when I use the provided docker [image](https://hub.docker.com/r/bitxiong/tsn/) to extract the optical_flow images using extract_optical_flow.sh the images are not created. The OUT_FOLDER gets filled with empty folders matching the structure of...

Wrong dimensions for Avg-Pool layer

2

I think there is a small mistake in the model of the i3d. https://github.com/tomrunia/PyTorchConv3D/blob/e6cc440e5613d6c826288402c66681a5e6ea5fa0/models/i3d.py#L303 In the paper QuVadis Action Recognition this is actuall 2,7,7 not 8,7,7.

Testing framework on THUMOS

10

@cinjon how is your progress to try this code on THUMOS? If I understand this correctly I have to extract the snippet level features using TSN (https://github.com/yjxiong/anet2016-cuhk). But the anet2016-cuhk...

Configure WandB run name in checkpoint path via Lightning CLI

3

### Description & Motivation It would be great to format the wandb run name: `wandb.run.name` into the checkpoint filename. epoch: 10 loss: 0.5423 filename: "epoch{epoch}-loss{val_loss:.2f}" --> "epoch10-loss0.54" filename: "{wandb.run.name}-{epoch}-{val_loss:.2f}" -->...

Is there a way to give data-pin-board-width a dynamic size? In the [Doku](https://developers.pinterest.com/docs/widgets/profile-widget/) it says that `data-pin-board-width` does on default `Fill width of parent` but this doesn't work for me....

Trace the Model to use in C++

1

With the Release of pytorch b1.0 it is now possible to load pytorch models trained in Python in C++. But for doing so it is necessary to trace the model...