Flownet-images file corrupt

[ae208gpu]

Hello @yaorong0921, I used the link for downloading the zip file of the flownet images and am not able to unzip the contents (using Mac). Below is the error I am getting please could you update the file and share the link asap

"unzip flownet2_road_camera.zip Archive: flownet2_road_camera.zip warning [flownet2_road_camera.zip]: 6454289820 extra bytes at beginning or within zipfile (attempting to process anyway) error [flownet2_road_camera.zip]: start of central directory not found; zipfile corrupt. (please check that you have transferred or created the zipfile in the appropriate BINARY mode and that you have compiled UnZip properly)"

[yaorong0921]

Hello @ae208gpu, I am using Windows and do not encounter any problems. Could you please try with a win system?

[ae208gpu]

@yaorong0921 the issue is with Outlook, refer this https://github.com/pmqs/Fix-OneDrive-Zip for fixing the zip after downloading

[ae208gpu]

@yaorong0921 do you have the code for getting the output by combining the inside and outside features, need to benchmark the model for a research work . Please would share the code for evaluating the model on inside-only, outside-only, and inside+outside

[ae208gpu]

Hello @yaorong0921 looking follow up on the request sent before: Please would share the code for evaluating the model on inside-only, outside-only, and inside+outside.

Please would you also share the inference script to view results like the .gif file?

[yaorong0921]

Hi @ae208gpu, I will try to find the script for combining since it was on a server that I am now having problems accessing. BTW, please refer to this explanation, if you would like to try to implement it yourself, which can be faster.

[ae208gpu]

@yaorong0921 I understand, but I am not able to generate any evaluation metric results using the scripts available that were stated in the paper. I am only looking to get those paper results by replicating the repo. Please would you provide support for that as it would be really helpful.

[Sidd1609]

@yaorong0921 which model was used to generate the flownet images ?

[yaorong0921]

@yaorong0921 which model was used to generate the flownet images ?

Hi @Sidd1609 , We used FlowNet 2.0 to extract the optical flow of all outside images.

[wasilone11]

@yaorong0921 how can we visualize the output results?

[yaorong0921]

Hi @wasilone11, for visualizing, I output results from the model for each test sample at each time step. Then I made the demo video with subtitles indicating the prediction (I just picked up one video from the test set for demo. To add subtitles, I used the video editor software.) I hope this information is helpful to you.

Best,