search

Lightning-Universe / Pose-app

A Lightning app for animal pose estimation.
https://pose-app.readthedocs.io/
Apache License 2.0
18 stars 1 forks source link

Handling Large SLEAP File (460MB) in the App #107

Open Freddie-Faulich opened 4 weeks ago

Freddie-Faulich commented 4 weeks ago

Hi everyone,

I'm working with a SLEAP file that is approximately 460MB in size. I’m wondering if there’s a way to effectively use such a large file within the app. Any advice or tips would be greatly appreciated.

Best regards

themattinthehatt commented 4 weeks ago

@Freddie-Faulich you can change the maximum upload size following the FAQ here: https://pose-app.readthedocs.io/en/latest/source/faqs.html (How do I increase the file upload size limit)

Freddie-Faulich commented 4 weeks ago

I tried that, but I can't find my labeled images anywhere

Shmuel-columbia commented 4 weeks ago

Hi @Freddie-Faulich, Can you please elaborate on what you are trying to do? Are you trying to export labeled images from SLEAP and import them to the app?

Freddie-Faulich commented 4 weeks ago

Yes that's exactly what I'm trying to do! Is this possible?

Shmuel-columbia commented 4 weeks ago

In your SLEAP project go to: predict > export labels package > labeled frames you should get an pkg.slp file, which you can import to Lightning Pose App by selecting: create new project from sourch > SLEAP. Launch the app as instructed in the FAQ link above if your file size is larger than 200MB. Here is a link to our docs for further instruction.

let us know if you have any questions. Here to help!

Freddie-Faulich commented 3 weeks ago

That's what I did, should my labeled data be visible anywhere in the GUI?

Shmuel-columbia commented 3 weeks ago

Yes. If the project was successfully uploaded, you should be able to see the labeled frames in the LABEL FRAMES tab. Log in to Label Studio (user and password in the first tab), and you should see the annotated frames there. OR, if you have a newer version of the app, in the EXTRACT FRAMES tab, there is the "Check Labels" feature, which imports the annotated frames and displays them in the app.

Freddie-Faulich commented 2 weeks ago

I use the app as intended but when i try to updoad my data the app stops unexpectedly, and I receive the following error:

INFO: Your Lightning App is being stopped. This won't take long. INFO: Your Lightning App has been stopped successfully! INFO: Received SIGTERM signal. Gracefully terminating train_ui.work... INFO: Received SIGTERM signal. Gracefully terminating streamlit_frame.work... INFO: Received SIGTERM signal. Gracefully terminating tensorboard... INFO: Received SIGTERM signal. Gracefully terminating label_studio.label_studio... INFO: Received SIGTERM signal. Gracefully terminating fiftyone... Traceback (most recent call last): File "/home/zeus/miniconda3/envs/cloudspace/bin/lightning_app", line 8, in sys.exit(_cli_entry_point()) ... File "/teamspace/studios/this_studio/Pose-app/lightning_pose_app/ui/train_infer.py", line 781, in _determine_dataset_type self.allow_context = is_context_dataset( File "/teamspace/studios/this_studio/Pose-app/lightning_pose_app/utilities.py", line 60, in is_context_dataset for frame in frames_in_dir: TypeError: iteration over a 0-d array INFO: Received SIGTERM signal. Gracefully terminating streamlit_video.work...

Could you please provide guidance on what might be causing this issue and how to resolve it? Best regards

Shmuel-columbia commented 2 weeks ago

Can you please attach the CollectedData.csv and the selected_frames.csv files? The CollectedData.csv should be in your main project folder and the selected_frames.csv should be in: labeled-data > name of your video > selected_frames.csv

themattinthehatt commented 5 days ago

@Freddie-Faulich just checking in on this again - do you see the files in your project folder that @Shmuel-columbia mentioned? (CollectedData.csv and labeled-data/<name_of_video>/selected_frames.csv)