-
Notifications
You must be signed in to change notification settings - Fork 54
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
running on multiple simultaneous videos #53
Comments
Hi Evan, did you ever figure this out? When I try to train on multiple h5 files (sleap data) my output sav files seem to only account for one h5 file. I'm running this in the API |
Hi Isabelle,
No, I never figured this out. But I would like to....
I managed to do runs training on up to 4 sessions and testing on 1 session, but I was never able to train and test on an arbitrary number of sessions (csv DLC outputs as inputs).
Do you have any ideas?
Thanks,
Evan
…________________________________
From: Isabelle ***@***.***>
Sent: Monday, March 25, 2024 7:02 AM
To: YttriLab/B-SOID ***@***.***>
Cc: Evan Vickers ***@***.***>; Author ***@***.***>
Subject: Re: [YttriLab/B-SOID] running on multiple simultaneous videos (Issue #53)
Hi Evan, did you ever figure this out? When I try to train on multiple h5 files (sleap data) my output sav files seem to only account for one h5 file. I'm running this in the API
—
Reply to this email directly, view it on GitHub<https://urldefense.com/v3/__https://github.com/YttriLab/B-SOID/issues/53*issuecomment-2018076380__;Iw!!C5qS4YX3!DEqBNAjk68tmu4zKxlpI6c3dMKjXkxl5gMqgjWrYI7-vW7Juu4UboPNGd3Agl1IqyFRHWH06vn3tA62YYS_tMqgpIOB2$>, or unsubscribe<https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AJGCDZUP6LXVPIFGCR46BUTY2AU7FAVCNFSM5GMO4RT2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TEMBRHAYDONRTHAYA__;!!C5qS4YX3!DEqBNAjk68tmu4zKxlpI6c3dMKjXkxl5gMqgjWrYI7-vW7Juu4UboPNGd3Agl1IqyFRHWH06vn3tA62YYS_tMqBBmQH5$>.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
Ah so are you running 4 separate sessions to train the random forest, but you can't cluster/train on multiple csv DLC outputs at once? I imagine it doesn't make sense to only get cluster results from one video and apply it to all future data.... this is what I'm having issues with too |
Have you tried using the scripts on github? I also saw they have google colab but I only could find the old t-sne version. |
To be honest it's been awhile since I looked at this.
I trained random forest on 4 together, but it failed when I tried to train on 5 or ~25. Then failed to get pose estimate output on more than one session at a time. Also failed when I tried to run one huge concatenated csv from 5 or 25 sessions. Also failed running github script on cluster computing resource or locally, in same cases.
Really no idea why, either csv formatting issue or data from many sessions too heterogeneous?
Keypoint moseq had no problem doing any of these things, but I would love to get a batch run of bsoid to run....or w/ asoid, which i installed and ran a successful test on but never really used.
Am gearing up to write my own code for behavior analysis, so plan to reopen this can of worms soon. : )
…-Evan
Sent from my T-Mobile 5G Device
Get Outlook for Android<https://aka.ms/AAb9ysg>
________________________________
From: Isabelle ***@***.***>
Sent: Monday, March 25, 2024 10:21:06 AM
To: YttriLab/B-SOID ***@***.***>
Cc: Evan Vickers ***@***.***>; Author ***@***.***>
Subject: Re: [YttriLab/B-SOID] running on multiple simultaneous videos (Issue #53)
Have you tried using the scripts on github? I also saw they have google colab but I only could find the old t-sne version.
—
Reply to this email directly, view it on GitHub<https://urldefense.com/v3/__https://github.com/YttriLab/B-SOID/issues/53*issuecomment-2018516055__;Iw!!C5qS4YX3!BU0UDZqGxQ_s2cDsbBRCEpnZN5zvlC0aHO5Zf9bJkKxdPZXozRsUEtO0bRLJ5h5_cqbq5O_IIrnBoCIZVvyeHumu8w6k$>, or unsubscribe<https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AJGCDZQSKH4HLLDYVZA2GBLY2BMIFAVCNFSM5GMO4RT2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TEMBRHA2TCNRQGU2Q__;!!C5qS4YX3!BU0UDZqGxQ_s2cDsbBRCEpnZN5zvlC0aHO5Zf9bJkKxdPZXozRsUEtO0bRLJ5h5_cqbq5O_IIrnBoCIZVvyeHlJ435Vk$>.
You are receiving this because you authored the thread.Message ID: ***@***.***>
|
huh do you remember how you trained on 4 together? I tried using the session 1 session 2 file structure they recommended and putting all h5 files with pose data into one folder (having an output folder too in both cases) and my results appeared to only represent one h5 file. I just downloaded A-SOID so fingers crossed that works but I also don't want to have to annotate manually...I'll have to look into moseq. Let me know how things go with your own code, best of luck! |
Hi,
Is it possible to run multiple simultaneously acquired videos with accompanying DLC pose estimate csv's using the GUI?
I would like to get composite states from the features of all 3 videos.
Thanks,
Evan
The text was updated successfully, but these errors were encountered: