Share your feedback with NiChart. It's quick and easy!

[gurayerus]

Hello,

The feedback from our community is incredibly valuable to NiChart. We would be grateful if you could take a few minutes to share your thoughts by completing our surveys.

• NiChart User Experience

• Shaping the Future of NiChart

You can view our quick-start cloud demo videos in our YouTube channel

• NiChart Demo1

• NiChart Demo2

Visit our web page for an overview of NiChart or simply run it on our cloud app

Thank you!

[A-Cronnex]

Hello, my name is Aaron. I am future researcher from the Univeristy of Panama. Two of the members of your project have personally recommended the usage of your pipeline, I have talked with two of them in the NITRC forum. However, I cannot use the tool in the same way as the tutorial shows because I can't choose the output directory based on my own data and neither I can choose the type of training I wanna do (In the DLMUSE part, I can't chose between cuda and the others) I am very eager to see if your tool can help me with my research project, but I don't know why it's missing such key options that appear in the tutorial.

Edit: I will share this comment also in the Software discussion part

[AlexanderGetka-cbica]

Hi Aaron, Thank you for your interest in NiChart. I'm happy to help troubleshoot. Are you using the cloud portal or the local version of the application? On the cloud, we don't allow users to select CUDA/MPS/etc. This option is only meant for local usage in case one has a Mac, doesn't have an Nvidia GPU, and so on (by default on cloud we always use CUDA). If this option isn't showing up on local, this may be a bug. Also on the cloud, output is automatically stored. To get results you can just click download at the end of each step (for instance, to download DLMUSE ROIs, scroll to the bottom after DLMUSE finishes, click download and select scans.) On the local version, you should be able to select the output folder, but if not you can check under the "output_folder" subdirectory in the cloned repo. For clarity, we don't currently offer training via this interface, only inference of our pretrained models on submitted data. I hope this is helpful, and please let me know if you have any questions. Best regards, Alex Getka Lead Application Developer Center for Biomedical Computing and Analytics University of Pennsylvania

…

________________________________ From: Aaron C. ***@***.***> Sent: Thursday, January 23, 2025 4:34 PM To: CBICA/NiChart_Project ***@***.***> Cc: Subscribed ***@***.***> Subject: [External] Re: [CBICA/NiChart_Project] Share your feedback with NiChart. It's quick and easy! (Discussion #72) Hello, my name is Aaron. I am future researcher from the Univeristy of Panama. Two of the members of your project have personally recommended the usage of your pipeline, I have talked with two of them in the NITRC forum. However, I cannot use the tool in the same way as the tutorial shows because I can't choose the output directory based on my own data and neither I can choose the type of training I wanna do (In the DLMUSE part, I can't chose between cuda and the others) I am very eager to see if your tool can help me with my research project, but I don't know why it's missing such key options that appear in the tutorial. — Reply to this email directly, view it on GitHub<#72 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AOHRPDQTN4LNTRHRYGQ5NDD2MFN75AVCNFSM6AAAAABT5MSTMKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOJTGUYTKOI>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[A-Cronnex]

Hi Alex, thank you! I appreciate your responde a lot

I am using the cloud version on a computer with Ubuntu. All the images are already in a Nifti format, so I am using directly the DLMUSE method.

I have already put the images under the DLMUSE. Is it taking quite a while, there are 39 images in total. Is it normal that it takes a lot of time, or I did something wrong in the process?

Edit: It worked correctly after a while of waiting

[AlexanderGetka-cbica]

Hi Aaron, Under normal conditions, in my observation, it typically takes perhaps around 20 to 40 seconds per T1 scan for the full DLMUSE pipeline. (We are always striving toward lowering this number and increasing parallel processing even further, but this is a complex process even on the cloud. Our first priority right now are changes that should speed up the process by a factor of approximately 10x or more, and this should be available within the next few months.) Looking through the logs right now, I see multiple submissions of 160+ scans — it seems the platform is a little more popular than we expected for this at the current moment, and this might be causing some issues with the processing speed due to high GPU usage. If your process hasn't finished, it may be best to try again later. Also note that, at least for up to 24 hours, you can simply let the process run and then download the results later. For example, if i create a study "MyFirstStudy", upload my files there, then run DLMUSE, I don't have to even stay on the page. If I go back to the same page later and select MyFirstStudy again, and results are available from the first run, it should still be able to be downloaded at the bottom. (However, this will not work if your results have been automatically cleaned up, which occurs after some time, or if the original run never finished). Hope this helps and I hope to get back to you soon regarding updates. Thanks, Alex

…

________________________________ From: Aaron C. ***@***.***> Sent: Thursday, January 23, 2025 7:53 PM To: CBICA/NiChart_Project ***@***.***> Cc: Getka, Alexander ***@***.***>; Comment ***@***.***> Subject: [External] Re: [CBICA/NiChart_Project] Share your feedback with NiChart. It's quick and easy! (Discussion #72) Hi Alex, thank you! I appreciate your responde a lot I am using the cloud version on a computer with Ubuntu. All the images are already in a Nifti format, so I am using directly the DLMUSE method. I have already put the images under the DLMUSE. Is it taking quite a while, there are 39 images in total. Is it normal that it takes a lot of time, or I did something wrong in the process? — Reply to this email directly, view it on GitHub<#72 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AOHRPDQIG32ZHKCF2DMX7WL2MGFHPAVCNFSM6AAAAABT5MSTMKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCOJTGY3DIMY>. You are receiving this because you commented.Message ID: ***@***.***>