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
Removing mictures of templates: VallueError operands could not be broadcast together #395
Comments
Hi, I'm working with Krzysztof to get SpyKING Cirus running. We've updated to the most recent Github version of SpyKING Circus, but we still encounter the same problem ( Additionally SpyKING Circus was not able to detect all of our files - it only sees one ( |
The second issue @mmagnuski mentioned was solved by commenting line 126 in neuralynx.py file and writing instead
The problem was caused by filter_name_duplicates function which strips names of files with "_" separator and removing last part of filename, so if your filename is NAME_number.ncs (in our case CSC_129), it will remove number and leave only NAME part so in theory you will have multiple copies of files named NAME to analyze and the program will take into account only first one. However, our main problem
|
You could try to add |
Yes, as @KKKenser mentions, we would have to modify the naming of our files to make it work in Spyking Circus without modifying the code. I think this could be considered a bug, because our original naming was in line with this documentation page. The files are not detected because Spying Circus detects duplicates by splitting file names by underscore and rejecting the last part - so |
However, while I know this is not entirely finished, and that the publication will only arrives in couple of month, you could try to give it a go to spyking-circus2, entirely based on SpikeInterface (https://github.com/SpikeInterface/). i'll write a proper documentation soon, but thanks to spikeinterface, the handling of the files/preprocessing, .... is much more smoother and easier. You can simply do somehting like
|
Hi @yger, thanks for your suggestion! We can try this out. Also, thank you for suggesting using spyking circus through spike_interface. Indeed, we will try giving it a go. I actually wanted to start with SC1, because I remember seeing figures in this paper that would suggest that SC2 is not as good as SC1 (for example number of correctly identified units in table 2). I don't know how accurate this table is with respect to current state to SC2, but this was the source of bias in our case. :) |
I would not trust the figure of this preprint, regarding the number obtained with SC2. It was clearly misused (not necessarily on purpose) and/or taken from a too alpha version that was not stabilized. Internal bencharmks are now clearly showing that SC2 and SC1 should behave alike. But you are welcome to try with SC1 first ! The best for you might be to share your data if you cant, in order for me to understand why is going on under the hood |
Great, thanks for reassuring us! |
We have received approval for sharing the data we've used for testing the spyking-circus environment. Here's the link to zip file on google disc: https://drive.google.com/file/d/1kpyRc11xCokAkPcPjyK--KcGcSS3TcHU/view?usp=sharing We'd be grateful if you could find the time and take a look at our issue |
We managed to run sorting without errors last week, but we are not sure what helped in the end. We will try again on our full dataset and let you know if we still have problems. So the issue can be closed - at least temporarily. |
Hi,
I'm trying to run test sorting on a few channels (NEURALYNX format) and I'm constantly getting following error:
And parameters file is following:
I have downloaded Spyking-Circus recently, so I'm assuming I have up-to-date version. With different initial settings in parameters file I have different values for brodcasting together ValueError (4), (8) or as above (2), (4). I have loaded so far only 8 channels for testing, in the mapping file radius covers all of them.
I would be grateful for any help in solving this issue.
Sincerly,
Krzysztof
The text was updated successfully, but these errors were encountered: