Refractory tests and corrected hh_cond_exp_traub (fixes #473) #590
Conversation
| "aeif_cond_beta_multisynapse", | ||
| "amat2_psc_exp", | ||
| "ginzburg_neuron", | ||
| "hh_cond_exp_traub", |
There was a problem hiding this comment.
Why do you need to ignore this model? Isn't that the one you are fixing in this PR?
There was a problem hiding this comment.
Yes, but the hh_* models do not clamp the potential, so I cannot easily check the refractory period... I'll try using a huge current
There was a problem hiding this comment.
No, I think there is just no way to check the HH models (I'm not even sure t_ref makes any sense for them)
| ignore_model = [ | ||
| "aeif_cond_alpha_RK5", | ||
| "aeif_cond_alpha_multisynapse", | ||
| "aeif_cond_beta_multisynapse", |
There was a problem hiding this comment.
Why do you need to exclude the multisynapse models?
There was a problem hiding this comment.
RK5's implementation of t_ref is still wrong and I was waiting for you to merge your PR regarding the other two ;)
Now that it's done, I'll include them in the tests tomorrow.
|
Ok, I corrected the test (and recorrected it when I remembered why I had built the list of neurons to be tested that way... -_-') |
|
@Silmathoron Very nice. But could you do one more update? The Hill-Tononi revision has just been merged this morning. Could you merge that into your branch and make sure that your test still works? |
|
This is what I did this morning just after abigail merged it ;) |
|
Sorry, my mistake. Github had listed all your recent commits under "... added some commits 2 days ago", so I thought it couldn't contain the revised HT. |
heplesser
added
ZC: Model
|
@jschuecker Could you be the second reviewer for this one? It is just a small (but important fix) and a new test. |
|
|
||
| # --------------------------------------------------------------------------- # | ||
| # Simulation and refractory time | ||
| # ------------------------- |
There was a problem hiding this comment.
you could change this to "refractory time limits" to avoid confusion as this are only the lower and upper bounds right?
There was a problem hiding this comment.
Done (and additional comments added)
| else: | ||
| Vr = nest.GetStatus(neuron, "V_reset")[0] | ||
| times = nest.GetStatus(vm, "events")[0]["times"] | ||
| idx_max = (np.argwhere(np.isclose(times, spike_times[1]))[0][0] |
There was a problem hiding this comment.
is there a reason why np.isclose is used here? Otherwise you could make this consistent with line 209.
There was a problem hiding this comment.
Yes, it's for precise models ;) I'll add a comment
EDIT: or maybe not anymore... I'll check that!
EDIT 2: corrected
|
@Silmathoron: Very nice new test! I only have two minor suggestions (see my inline comments). |
| @@ -134,14 +134,14 @@ | |||
|
|
|||
|
|
|||
| # --------------------------------------------------------------------------- # | |||
| # Simulation and refractory time | |||
| # Simulation and refractory time limits | |||
There was a problem hiding this comment.
@Silmathoron, Sorry, I meant to change this to "Simulation time and refractory time limits"
|
@Silmathoron thanks for the corrections. 👍 from my side |
This PR adresses #473 by correcting the refractory behaviour of
hh_cond_exp_traubIt also implements a random refractory test for many hybrid models.