Closed jgosmann closed 7 years ago
Thank you! I'm pretty sure I know how to fix this (it's a dumb bug on my part where I didn't check if something is an ObjView), but in order to test it, I need to be able to run your model. Could you add in the needed import statements? I'm having trouble finding where some of those things are....
It's the tcm
branch of this repository. Run pytry imem/models.py --gui
. The problematic network is net_m_tf
(and probably net_m_ft
too) in the tcm
network. You will need nengo_spa
to run it (and you might have to pip install the repository).
It's the tcm branch of [this repository].
Hmm, there doesn't seem to be a tcm branch in this repository.......
Sorry, I forgot to insert the actual link.
Sorry, I forgot to insert the actual link.
Hmm, that doesn't seem to run for me... Do I need a particular nengo version?
running NeuralTCM#20170518-122541-ed06ce50
Traceback (most recent call last):
File "C:\Python27\Scripts\pytry-script.py", line 9, in <module>
load_entry_point('pytry', 'console_scripts', 'pytry')()
File "c:\users\terry\documents\github\pytry\pytry\cmdline.py", line 44, in run
t.run(**args)
File "c:\users\terry\documents\github\pytry\pytry\trial.py", line 88, in run
result = self.execute_trial(p)
File "c:\users\terry\documents\github\pytry\pytry\nengo.py", line 48, in execute_trial
model = self.model(p)
File ".\imem\models.py", line 38, in model
model.tcm = tcm.TCM(p.beta, self.control, p.item_d, p.context_d)
File "c:\users\terry\documents\github\imem\imem\tcm.py", line 285, in __init__
self.context_vocab, self.item_vocab)
File "c:\users\terry\documents\github\imem\imem\tcm.py", line 148, in __init__
self.target.output[start:end], conn.learning_rule)
File "c:\users\terry\documents\github\nengo\nengo\base.py", line 35, in __call__
inst.__init__(*args, **kwargs)
File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 424, in __init__
self.post = post
File "c:\users\terry\documents\github\nengo\nengo\base.py", line 106, in __setattr__
super(NengoObject, self).__setattr__(name, val)
File "c:\users\terry\documents\github\nengo\nengo\config.py", line 454, in __setattr__
super(SupportDefaultsMixin, self).__setattr__(name, val)
File "c:\users\terry\documents\github\nengo\nengo\params.py", line 112, in __set__
new_value = self.validate(instance, value)
File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 31, in validate
return super(PrePostParam, self).validate(conn, nengo_obj)
File "c:\users\terry\documents\github\nengo\nengo\base.py", line 220, in validate
if self.nonzero_size_in and nengo_obj.size_in < 1:
File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 624, in size_in
return getattr(self, self.default_error_signal).size_in
File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 617, in default_error_signal
"Directly connecting to the learning rule is only allowed if"
ValueError: Directly connecting to the learning rule is only allowed ifit uses exactly one error signal.
And actually, since this is such a small fix, could you check if this commit fixes it for you?
https://github.com/nengo/nengo_gui/commit/393e8bb16f13a5a25a56e93fad10ab3a157e0003
One of nengo master or most recent release should work. I can test it myself once I got my computer working again
On May 18, 2017 12:32:11 PM EDT, tcstewar notifications@github.com wrote:
Sorry, I forgot to insert the actual link.
Hmm, that doesn't seem to run for me... Do I need a particular nengo version?
running NeuralTCM#20170518-122541-ed06ce50 Traceback (most recent call last): File "C:\Python27\Scripts\pytry-script.py", line 9, in <module> load_entry_point('pytry', 'console_scripts', 'pytry')() File "c:\users\terry\documents\github\pytry\pytry\cmdline.py", line 44, in run t.run(**args) File "c:\users\terry\documents\github\pytry\pytry\trial.py", line 88, in run result = self.execute_trial(p) File "c:\users\terry\documents\github\pytry\pytry\nengo.py", line 48, in execute_trial model = self.model(p) File ".\imem\models.py", line 38, in model model.tcm = tcm.TCM(p.beta, self.control, p.item_d, p.context_d) File "c:\users\terry\documents\github\imem\imem\tcm.py", line 285, in __init__ self.context_vocab, self.item_vocab) File "c:\users\terry\documents\github\imem\imem\tcm.py", line 148, in __init__ self.target.output[start:end], conn.learning_rule) File "c:\users\terry\documents\github\nengo\nengo\base.py", line 35, in __call__ inst.__init__(*args, **kwargs) File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 424, in __init__ self.post = post File "c:\users\terry\documents\github\nengo\nengo\base.py", line 106, in __setattr__ super(NengoObject, self).__setattr__(name, val) File "c:\users\terry\documents\github\nengo\nengo\config.py", line 454, in __setattr__ super(SupportDefaultsMixin, self).__setattr__(name, val) File "c:\users\terry\documents\github\nengo\nengo\params.py", line 112, in __set__ new_value = self.validate(instance, value) File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 31, in validate return super(PrePostParam, self).validate(conn, nengo_obj) File "c:\users\terry\documents\github\nengo\nengo\base.py", line 220, in validate if self.nonzero_size_in and nengo_obj.size_in < 1: File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 624, in size_in return getattr(self, self.default_error_signal).size_in File "c:\users\terry\documents\github\nengo\nengo\connection.py", line 617, in default_error_signal "Directly connecting to the learning rule is only allowed if" ValueError: Directly connecting to the learning rule is only allowed ifit uses exactly one error signal.
And actually, since this is such a small fix, could you check if this commit fixes it for you?
https://github.com/nengo/nengo_gui/commit/393e8bb16f13a5a25a56e93fad10ab3a157e0003
-- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: https://github.com/nengo/nengo_gui/issues/893#issuecomment-302463842
-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
One of nengo master or most recent release should work. I can test it myself once I got my computer working again
Looks like there's a bunch of Python3 vs Python2 issues causing the problem, which means I can't test it from home. But I'm 99% sure that fix above will do it for you.
I can verify the fix works.
The network causing this error: