You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
note: This appears to be a different error than issue #22 (different error, different line in the script it seems). tried a solution mentioned there but didn't work.
So for on one dataset I have I get the following error at the extracting waveform part:
Extracting waveforms : 0%| | 0/521 [00:00<?, ?it/s]/home/rdorman/miniconda3/envs/klusta/lib/python3.5/site-packages/klusta/traces/waveform.py:94: RuntimeWarning: divide by zero encountered in true_divide
return np.clip((x - tw) / (ts - tw), 0, 1)
/home/rdorman/miniconda3/envs/klusta/lib/python3.5/site-packages/klusta/traces/waveform.py:94: RuntimeWarning: invalid value encountered in true_divide
return np.clip((x - tw) / (ts - tw), 0, 1)
ValueError: cannot convert float NaN to integer
Seems clear, there appear to be something with nans and dividing by zero. But, similar recordings work just fine, and in all other ways this recording doesn't seem to be different (i.e. raw traces and MUA script all work fine, there are no nans in the data, and no other strange things to be seen).
In the waveforms.py, these variables tw and ts are weak threshold and strong threshold, respectively. These are 2 and 4 (same as in the clustering for all my other recordings). But, i'm not fluent in python so I might miss something here.
I'm kind of in the dark here. No idea why it happens with just one recording (which seems fine otherwise). Any comments or help are appreciated.
Cheers, RD
The text was updated successfully, but these errors were encountered:
Error seems to have dissappears, the only thing I recall that happened in between was an update of the code. Haven't seen the error since. Will update should the error return
edit 9/4: error happends again, consequently on the same data
note: This appears to be a different error than issue #22 (different error, different line in the script it seems). tried a solution mentioned there but didn't work.
So for on one dataset I have I get the following error at the extracting waveform part:
Extracting waveforms : 0%| | 0/521 [00:00<?, ?it/s]/home/rdorman/miniconda3/envs/klusta/lib/python3.5/site-packages/klusta/traces/waveform.py:94: RuntimeWarning: divide by zero encountered in true_divide
return np.clip((x - tw) / (ts - tw), 0, 1)
/home/rdorman/miniconda3/envs/klusta/lib/python3.5/site-packages/klusta/traces/waveform.py:94: RuntimeWarning: invalid value encountered in true_divide
return np.clip((x - tw) / (ts - tw), 0, 1)
ValueError: cannot convert float NaN to integer
Seems clear, there appear to be something with nans and dividing by zero. But, similar recordings work just fine, and in all other ways this recording doesn't seem to be different (i.e. raw traces and MUA script all work fine, there are no nans in the data, and no other strange things to be seen).
In the waveforms.py, these variables tw and ts are weak threshold and strong threshold, respectively. These are 2 and 4 (same as in the clustering for all my other recordings). But, i'm not fluent in python so I might miss something here.
I'm kind of in the dark here. No idea why it happens with just one recording (which seems fine otherwise). Any comments or help are appreciated.
Cheers, RD
The text was updated successfully, but these errors were encountered: