Preserve the mean-field weight parameter during normalization by AHMETHAKANBEZIR1 · Pull Request #813 · BindsNET/bindsnet

AHMETHAKANBEZIR1 · 2026-10-21T13:11:22Z

MeanFieldConnection.normalize replaces its registered Parameter with a reshaped tensor and crashes for shared and per-target weights. Scale the existing Parameter in place, preserving its identity, shape and state_dict registration.

Apply the normalization requested in review: norm is the total absolute weight across target neurons. Divide by the sum of absolute values to preserve inhibitory signs and prevent cancellation, and count a shared scalar once per target neuron. With three targets and norm=2, both a shared weight of 0.25 and the equivalent per-target vector normalize to 2/3 at every target. Zero weights remain finite. Constructor and method documentation now use this convention.

The branch includes master 78ec02b, including the merged #814 kernel-normalization change; the changelog conflict was resolved with all entries retained.

Validation on Windows CPU, Python 3.12 / Torch 2.14.0:

The updated review cases fail 10 times on the preceding implementation, with 17 passing controls; all 27 cases pass after the revision.
Complete standalone suite: 282 passed, 1 skipped. Coverage includes negative and cancelling weights, shared/per-target compute equivalence, repeated normalization, zero and disabled norms, Parameter identity and Network.run.
Private integration with Reset and resize CSRM connection spike history #815 and Forward mean-field learning rule options by name #816: 297 passed, 1 skipped. The changes applied cleanly apart from retaining each changelog entry. This is compatibility evidence, not a requirement to merge the PRs together.
Configured isort/Black hooks and diff checks passed. The five current-head GitHub checks succeeded. CUDA and timing measurements were not run locally.

AI disclosure: OpenAI Codex autonomously prepared and tested this contribution, including the requested revision. The account owner has not independently reviewed the code. No personal production incident or task-accuracy improvement is claimed.

Hananel-Hazan

Hi AhmetHakanBezir,

Thank you for using BindsNET and for opening this PR. The crash fix is right: on master, normalize() fails in every case I tried, and with your change the weight stays the same registered parameter. A couple of things need to change before it can merge. I measured each of them on your branch:

Negative weights flip sign. The code divides by the plain sum of the weights, and the sum of negative weights is negative. With norm=2, weights [-1, -2] become [0.67, 1.33]. With norm=1, a shared weight of -0.5 becomes 1.0. An inhibitory connection silently becomes an excitatory one.
Weights become infinite when their signs cancel. [1, -1] becomes [inf, -inf], and [1, -0.999999] becomes about [986895, -986894]. The guard only handles weights that are all zero.
A shared weight and per-neuron weights give different results. For 3 target neurons and norm=2, a shared weight of 0.25 becomes 2.0, but per-neuron weights [0.25, 0.25, 0.25] become 0.67 each. That is the same network, with three times more input in one case.

Please divide by the sum of the absolute values, which is what Connection.normalize already does, and count a shared weight once for each target neuron:

if self.norm is not None:
    # A scalar weight is shared by every target neuron.
    total = self.w.abs().sum() * (self.target.n / self.w.numel())
    self.w *= self.norm / total.masked_fill(total == 0, 1)

I tried this on your branch:

signs stay the same ([-1, -2] becomes [-0.67, -1.33]);
[1, -1] stays [1, -1];
the shared and per-neuron cases now agree at 0.67;
it runs on GPU with no extra synchronization, at the same speed as your version.

Your three shared-weight tests expect norm itself and will need the new expected values.

Also:

Add tests for negative weights, for mixed-sign weights that sum to zero, and for a shared weight compared with an equal per-neuron vector.
Update the norm description in __init__, which still says "Total weight per target neuron", so it matches the new docstring.
Rebase on master when it's ready; the changelog entry will need a small conflict fix.

Co-authored-by: OpenAI Codex <codex@openai.com>

AHMETHAKANBEZIR1 · 2026-10-23T14:47:08Z

Implemented in 53dd2e3: normalization uses the absolute weight total, counts a shared weight once per target neuron, preserves inhibitory signs, and keeps zero/cancelling weights finite. The constructor documentation and regression expectations are updated, and the branch includes master 78ec02b.

All 27 mean-field tests pass; the full CPU suite has 282 passes and 1 skip. All five current-head CI checks are green. CUDA and timing were not measured locally.

…, CODEOWNERS (#828) * test: golden-output and invariant regression tests, all-class smoke tests, API snapshot, coverage floor, PR template Phases 1-4, 6, 7 of the test hardening plan (logs/test_hardening_2026-10/PLAN.md). - test/regression/: 16 seeded networks ported from the review canaries; golden spike/voltage/trace/weight files (spikes exact, floats atol 1e-6, skipped on a different torch major.minor); invariants (split run, reset vs fresh, state_dict round trip, learning off, batch of copies, dt scaling). 100 passed, 15 strict xfail for known bugs (rule state outside state_dict, MulticompartmentConnection weights not saved, CSRM s_w window, PR #813 and #815 crashes). - test/network/test_all_classes.py: every Nodes, connection, learning rule, feature and model class built and run; 25 strict xfail pin known crashes. Core coverage network 79-90% -> 91%, learning 84-87% -> 95%, models 43% -> 99%. - test/test_api_surface.py + api_surface.json: public signatures pinned; regen scripts refuse to run without --reason. - CI: pytest-cov in dev group, coverage floor script (network 90, learning 94), coverage.xml artifact. - CONTRIBUTING.md, PR template, test/regression/README.md, bindsnet/learning/README.md known limitations. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * test: operation-count tests and A/B timing workflow (Phase 5) - test/regression/test_op_counts.py + op_counts.json: kernel dispatch counts for one update() of every learning rule (batch 1 and 4), one DiehlAndCook2015 step (batch 1 and 16) and LIFNodes.forward; fails when a count grows, warns on a drop. regen_op_counts.py refuses without --reason. 32 tests, 3.9 s. - .github/workflows/bench.yml + scripts/bench_ab.py: non-required PR workflow, merge base vs head, alternating rounds, median ratio with bootstrap interval, SLOWER above 1.25; report in the job summary and an artifact. Local no-change run: medians 0.965 to 1.012; a 0.5 ms/step sleep flagged on all workloads. Calibration on GitHub runners still pending. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> * ci: test guard (fails-before-fix and weakened-tests checks) and CODEOWNERS Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * ci: weakened-tests also checks helper functions and tolerance constants; document the guard checks Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * ci: coverage floors from the first GitHub CI measurement (network 90.6%, learning 95.2%) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>

* fix: IzhikevichNodes and CSRMNodes accept multi-dimensional shapes; CSRMNodes reset clears its spike window IzhikevichNodes sized its parameters by the n argument (None when only shape is given) and never reshaped them; CSRMNodes contracted with 'i,kij->kj'. Both now work for any shape; one-dimensional results are bit-identical (goldens conv2d_if_izh and csrm_window trial 0 unchanged). CSRMNodes.reset_state_variables now zeroes last_spikes, which changes results after a reset (csrm_window golden regenerated with that reason). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: LocalConnection batch > 1; MaxPool connections without decay or with one channel LocalConnection.compute dropped the batch dimension in its final view. MaxPool{1,2,3}d multiplied a missing decay (None) and squeeze()d away a channel of size 1; a missing decay now means plain spike counts, and the firing-rate buffer follows the spike tensor's shape. Results that ran before are unchanged (goldens pass). Reproducers kept as test_fixed_bug_stays_fixed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: sparse MulticompartmentConnection learns and runs with batch > 1; NetworkMonitor reset; tuple target shapes - MCC_LearningRule bounded sparse weights into a new tensor that only the rule held, so DiehlAndCook2015(sparse=True) never learned; the bound and the sparse normalisation now write in place. At batch 1 sparse and dense learn bit-identical weights without norm (test_sparse_dc2015_learns_like_dense); with norm they differ by summation order (max 6.0e-8). - MulticompartmentConnection.compute applies per-sample sparse weight copies ([batch, src, tgt]) with bmm; one copy keeps the old arithmetic. - NetworkMonitor.reset_state_variables looked connections up in layers. - [batch] + target.shape failed for tuple shapes. Results change: DiehlAndCook2015(sparse=True) now learns. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MSTDP/MSTDPET/Rmax state cleared on reset; Rmax and MCC MSTDPET work with batch > 1 - LearningRule.reset_state_variables deletes the tensors a rule creates on its first update (_step_state); AbstractConnection.reset_state_variables forwards to the rule. Reset then equals a freshly built network (invariant tests for mstdp_mstdpet and rules_mix now pass). Results change after a reset with learning on: goldens mstdp_mstdpet(_dt05), rules_mix(_dt05) regenerated, trial 0 bit-identical. - Rmax and MCC_learning.MSTDPET keep per-sample state and reduce the update over the batch, as MSTDP does. Batch 1 is bit-identical to the old formula (TestMccMstdpetBatched, goldens); batch 2 with sum equals the sum of single samples to summation order. Op counts at batch 1: MCC MSTDPET 24 -> 27, Rmax 24 -> 26 (regenerated with reason). API snapshot: +1 method. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: preserve the mean-field weight parameter during normalization (#813) Adopted from #813 (head 53dd2e3) by AHMETHAKANBEZIR1, reviewed with the bindsnet-pr-review skill (sandboxed canaries, fail-then-pass, API, timing); Hananel adopted it as accountable author because the PR says no human reviewed it. MeanFieldConnection.normalize scales w in place by norm over the absolute weight sum (a shared scalar counted once per target), instead of assigning a plain tensor to the Parameter (TypeError on every call). Co-Authored-By: Hananel Hazan <hananel@hazan.org.il> Co-Authored-By: OpenAI Codex <codex@openai.com> * fix: reset and resize the CSRM connection spike history (#815) Adopted from #815 (head 49c57f8) by AHMETHAKANBEZIR1, reviewed with the bindsnet-pr-review skill; Hananel adopted it as accountable author because the PR says no human reviewed it. Connection.compute_window re-creates s_w when the batch size changes, and reset_state_variables clears it for CSRM targets. Co-Authored-By: Hananel Hazan <hananel@hazan.org.il> Co-Authored-By: OpenAI Codex <codex@openai.com> * test: goldens and invariants after adopting #813 and #815 meanfield_norm and csrm_window goldens regenerated (both used to record a crash). Invariant xfails removed: reset-equals-fresh for csrm_window and meanfield_norm, round trip and batch for meanfield_norm. Still open: the CSRM spike window is not in the state_dict (round trip). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: CSRM spike window saved in the state_dict; MulticompartmentConnection drives CSRMNodes - Connection.s_w is a registered buffer, so state_dict saves it; a load adopts a saved window or drops one the saved state lacked (_load_window). The csrm_window round-trip invariant now passes; golden regenerated (adds s_w, no value changed). - MulticompartmentConnection.compute_window read s_w and self.w, neither of which it has, so MCC into CSRMNodes always raised. It now keeps a window of its own outputs (pipeline runs once per step), resized on a batch change, cleared on reset, saved in the state_dict. With static weights it matches a Connection bit for bit (test_multicompartment_into_csrm_matches_connection). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MeanFieldConnection keeps batch samples apart and applies weight_decay - compute() took the mean over the whole batch, so one sample's input depended on the others; it now uses each sample's own mean (batch 1 bit-identical: goldens pass without the weight_decay change). - weight_decay was passed by position into AbstractConnection's reduction slot and ignored; it is now passed by name. Results change for anyone who set weight_decay on a MeanFieldConnection (golden meanfield_decay regenerated with that reason). - Two #813 test lines expected the batch-wide mean (0.75); corrected to the per-sample means. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MeanField feature, Normalization/Updating sub-features, adaptation features at batch > 1, sparse weight bounds - topology_features.MeanField never called the base constructor (no name) and had no initial value; it now has an optional name, a placeholder value, and a per-sample mean ([batch, src, tgt], applied per sample by compute). - Sub-features (Normalization, Updating) get no-op prime_feature, update, normalize and reset_state_variables, which MulticompartmentConnection calls on every pipeline entry; both now work in the constructor and in Network.run. - AdaptationBaseSynapsHistory/OtherSynaps: a synapse is active when its pre-synaptic neuron spiked in any sample (the mask is shared connectivity); batch 1 gives the same values. - Sparse weights with wmin/wmax: Connection.__init__ bounds a dense copy, LearningRule.update bounds in place via a dense copy, and WeightDependentPostPre/DiehlAndCook compute their factors from a dense view. PostPre, WeightDependentPostPre and DiehlAndCook on a bounded SparseConnection learn bit-identical weights to the dense Connection. - Test workarounds for these bugs removed (batch-1-only sets, broken pairs, tuple-shape and sparse-bound special cases). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: known-bug list becomes FIXED_BUGS; workarounds removed; MulticompartmentConnection feature values in the state_dict - MulticompartmentConnection saves its pipeline feature values as feature.<name> and loads them in place, so DiehlAndCook2015 can be saved and restored (the last four strict xfails in test_invariants.py now pass; dc2015 goldens regenerated: new keys, no value changed). - Every KNOWN_BUGS entry is fixed: the dict becomes FIXED_BUGS (reproducers that must keep running); the empty xfail machinery in test_invariants.py and the batch-1, no-learning-check and sub-feature workarounds in test_all_classes.py are removed, so those classes run at batch 2 with full checks. - CONTRIBUTING, PR template and bindsnet/learning/README.md updated. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: CHANGELOG for the 0.3.5 bug fixes, results changes listed first Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: issues found by independent review of PR #831 Network (reviewer B): - MulticompartmentConnection state_dict: refuses to save duplicate feature names (they shared one key, and loading put one feature's value into the other); checks shape and sparse layout on load (copy_ broadcast a wrong-shaped tensor silently); saves the adaptation features' mask, spike buffer and counter (only value was saved). - An Updating sub-feature drives its parent's learning; the connection's update skips that parent (it learned twice per step). - A sub-feature whose parent is not in the pipeline is refused at construction (it failed later with an unclear TypeError). insert_pipeline had list.insert's arguments swapped. - Connection.compute_window and MaxPool resets create state on the spikes' device (CUDA device mismatch, also on master). - Docstrings state how the two CSRM windows differ once weights learn. Learning rules (reviewer A): - MCC_learning.DiehlAndCook crashed on sparse weights (twin of the learning.py fix). Sparse MCC weight copies now all get the batch's shared update in PostPre, DiehlAndCook and MSTDPET (as MSTDP did), so sparse equals dense; the default squeeze reduction means a sum over a larger batch, as in the fused dense path (_reduce_batch, _add_to_feature). - Docs: per-sample reward shape [batch, 1, 1]; reset keeps the reduction chosen at construction. CHANGELOG: checkpoint compatibility (strict=False for old checkpoints), new state shapes, sparse and Updating semantics. Each fix has a test that fails on ce98f93 and passes here. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: AHMETHAKANBEZIR1 <ahmethakan0008@gmail.com> Co-authored-by: OpenAI Codex <codex@openai.com>

* fix: IzhikevichNodes and CSRMNodes accept multi-dimensional shapes; CSRMNodes reset clears its spike window IzhikevichNodes sized its parameters by the n argument (None when only shape is given) and never reshaped them; CSRMNodes contracted with 'i,kij->kj'. Both now work for any shape; one-dimensional results are bit-identical (goldens conv2d_if_izh and csrm_window trial 0 unchanged). CSRMNodes.reset_state_variables now zeroes last_spikes, which changes results after a reset (csrm_window golden regenerated with that reason). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: LocalConnection batch > 1; MaxPool connections without decay or with one channel LocalConnection.compute dropped the batch dimension in its final view. MaxPool{1,2,3}d multiplied a missing decay (None) and squeeze()d away a channel of size 1; a missing decay now means plain spike counts, and the firing-rate buffer follows the spike tensor's shape. Results that ran before are unchanged (goldens pass). Reproducers kept as test_fixed_bug_stays_fixed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: sparse MulticompartmentConnection learns and runs with batch > 1; NetworkMonitor reset; tuple target shapes - MCC_LearningRule bounded sparse weights into a new tensor that only the rule held, so DiehlAndCook2015(sparse=True) never learned; the bound and the sparse normalisation now write in place. At batch 1 sparse and dense learn bit-identical weights without norm (test_sparse_dc2015_learns_like_dense); with norm they differ by summation order (max 6.0e-8). - MulticompartmentConnection.compute applies per-sample sparse weight copies ([batch, src, tgt]) with bmm; one copy keeps the old arithmetic. - NetworkMonitor.reset_state_variables looked connections up in layers. - [batch] + target.shape failed for tuple shapes. Results change: DiehlAndCook2015(sparse=True) now learns. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MSTDP/MSTDPET/Rmax state cleared on reset; Rmax and MCC MSTDPET work with batch > 1 - LearningRule.reset_state_variables deletes the tensors a rule creates on its first update (_step_state); AbstractConnection.reset_state_variables forwards to the rule. Reset then equals a freshly built network (invariant tests for mstdp_mstdpet and rules_mix now pass). Results change after a reset with learning on: goldens mstdp_mstdpet(_dt05), rules_mix(_dt05) regenerated, trial 0 bit-identical. - Rmax and MCC_learning.MSTDPET keep per-sample state and reduce the update over the batch, as MSTDP does. Batch 1 is bit-identical to the old formula (TestMccMstdpetBatched, goldens); batch 2 with sum equals the sum of single samples to summation order. Op counts at batch 1: MCC MSTDPET 24 -> 27, Rmax 24 -> 26 (regenerated with reason). API snapshot: +1 method. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: preserve the mean-field weight parameter during normalization (#813) Adopted from #813 (head 53dd2e3) by AHMETHAKANBEZIR1, reviewed with the bindsnet-pr-review skill (sandboxed canaries, fail-then-pass, API, timing); Hananel adopted it as accountable author because the PR says no human reviewed it. MeanFieldConnection.normalize scales w in place by norm over the absolute weight sum (a shared scalar counted once per target), instead of assigning a plain tensor to the Parameter (TypeError on every call). Co-Authored-By: Hananel Hazan <hananel@hazan.org.il> Co-Authored-By: OpenAI Codex <codex@openai.com> * fix: reset and resize the CSRM connection spike history (#815) Adopted from #815 (head 49c57f8) by AHMETHAKANBEZIR1, reviewed with the bindsnet-pr-review skill; Hananel adopted it as accountable author because the PR says no human reviewed it. Connection.compute_window re-creates s_w when the batch size changes, and reset_state_variables clears it for CSRM targets. Co-Authored-By: Hananel Hazan <hananel@hazan.org.il> Co-Authored-By: OpenAI Codex <codex@openai.com> * test: goldens and invariants after adopting #813 and #815 meanfield_norm and csrm_window goldens regenerated (both used to record a crash). Invariant xfails removed: reset-equals-fresh for csrm_window and meanfield_norm, round trip and batch for meanfield_norm. Still open: the CSRM spike window is not in the state_dict (round trip). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: CSRM spike window saved in the state_dict; MulticompartmentConnection drives CSRMNodes - Connection.s_w is a registered buffer, so state_dict saves it; a load adopts a saved window or drops one the saved state lacked (_load_window). The csrm_window round-trip invariant now passes; golden regenerated (adds s_w, no value changed). - MulticompartmentConnection.compute_window read s_w and self.w, neither of which it has, so MCC into CSRMNodes always raised. It now keeps a window of its own outputs (pipeline runs once per step), resized on a batch change, cleared on reset, saved in the state_dict. With static weights it matches a Connection bit for bit (test_multicompartment_into_csrm_matches_connection). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MeanFieldConnection keeps batch samples apart and applies weight_decay - compute() took the mean over the whole batch, so one sample's input depended on the others; it now uses each sample's own mean (batch 1 bit-identical: goldens pass without the weight_decay change). - weight_decay was passed by position into AbstractConnection's reduction slot and ignored; it is now passed by name. Results change for anyone who set weight_decay on a MeanFieldConnection (golden meanfield_decay regenerated with that reason). - Two #813 test lines expected the batch-wide mean (0.75); corrected to the per-sample means. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: MeanField feature, Normalization/Updating sub-features, adaptation features at batch > 1, sparse weight bounds - topology_features.MeanField never called the base constructor (no name) and had no initial value; it now has an optional name, a placeholder value, and a per-sample mean ([batch, src, tgt], applied per sample by compute). - Sub-features (Normalization, Updating) get no-op prime_feature, update, normalize and reset_state_variables, which MulticompartmentConnection calls on every pipeline entry; both now work in the constructor and in Network.run. - AdaptationBaseSynapsHistory/OtherSynaps: a synapse is active when its pre-synaptic neuron spiked in any sample (the mask is shared connectivity); batch 1 gives the same values. - Sparse weights with wmin/wmax: Connection.__init__ bounds a dense copy, LearningRule.update bounds in place via a dense copy, and WeightDependentPostPre/DiehlAndCook compute their factors from a dense view. PostPre, WeightDependentPostPre and DiehlAndCook on a bounded SparseConnection learn bit-identical weights to the dense Connection. - Test workarounds for these bugs removed (batch-1-only sets, broken pairs, tuple-shape and sparse-bound special cases). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: known-bug list becomes FIXED_BUGS; workarounds removed; MulticompartmentConnection feature values in the state_dict - MulticompartmentConnection saves its pipeline feature values as feature.<name> and loads them in place, so DiehlAndCook2015 can be saved and restored (the last four strict xfails in test_invariants.py now pass; dc2015 goldens regenerated: new keys, no value changed). - Every KNOWN_BUGS entry is fixed: the dict becomes FIXED_BUGS (reproducers that must keep running); the empty xfail machinery in test_invariants.py and the batch-1, no-learning-check and sub-feature workarounds in test_all_classes.py are removed, so those classes run at batch 2 with full checks. - CONTRIBUTING, PR template and bindsnet/learning/README.md updated. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs: CHANGELOG for the 0.3.5 bug fixes, results changes listed first Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix: issues found by independent review of PR #831 Network (reviewer B): - MulticompartmentConnection state_dict: refuses to save duplicate feature names (they shared one key, and loading put one feature's value into the other); checks shape and sparse layout on load (copy_ broadcast a wrong-shaped tensor silently); saves the adaptation features' mask, spike buffer and counter (only value was saved). - An Updating sub-feature drives its parent's learning; the connection's update skips that parent (it learned twice per step). - A sub-feature whose parent is not in the pipeline is refused at construction (it failed later with an unclear TypeError). insert_pipeline had list.insert's arguments swapped. - Connection.compute_window and MaxPool resets create state on the spikes' device (CUDA device mismatch, also on master). - Docstrings state how the two CSRM windows differ once weights learn. Learning rules (reviewer A): - MCC_learning.DiehlAndCook crashed on sparse weights (twin of the learning.py fix). Sparse MCC weight copies now all get the batch's shared update in PostPre, DiehlAndCook and MSTDPET (as MSTDP did), so sparse equals dense; the default squeeze reduction means a sum over a larger batch, as in the fused dense path (_reduce_batch, _add_to_feature). - Docs: per-sample reward shape [batch, 1, 1]; reset keeps the reduction chosen at construction. CHANGELOG: checkpoint compatibility (strict=False for old checkpoints), new state shapes, sparse and Updating semantics. Each fix has a test that fails on ce98f93 and passes here. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * ci: close test-guard bypasses found by independent review (WIP part 1) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test: op-count workloads for every Nodes class Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * ci: close the test-guard bypasses found by independent review (part 2), docs Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: AHMETHAKANBEZIR1 <ahmethakan0008@gmail.com> Co-authored-by: OpenAI Codex <codex@openai.com>

Hananel-Hazan · 2026-10-29T22:01:28Z

Thank you. This fix is now in master as part of #831 (commit d48dc79), which fixed all known bugs for 0.3.5. You are credited as co-author on that commit. The fix will ship in release 0.3.5. Closing, because the change is merged through #831.

AHMETHAKANBEZIR1 mentioned this pull request Oct 21, 2026

Keep zero convolutional and local kernels finite during normalization #814

Merged

Hananel-Hazan requested changes Oct 23, 2026

View reviewed changes

AHMETHAKANBEZIR1 and others added 3 commits October 23, 2026 04:47

fix: preserve mean-field weights during normalization

67573d0

Co-authored-by: OpenAI Codex <codex@openai.com>

fix: keep zero mean-field weights finite on repeated normalization

85709ff

Co-authored-by: OpenAI Codex <codex@openai.com>

fix: normalize mean-field absolute weights across targets

53dd2e3

Co-authored-by: OpenAI Codex <codex@openai.com>

AHMETHAKANBEZIR1 force-pushed the fix/meanfield-normalization-parameter branch from 4025c6a to 53dd2e3 Compare October 23, 2026 01:50

This was referenced Oct 23, 2026

Reset and resize CSRM connection spike history #815

Closed

Forward mean-field learning rule options by name #816

Closed

Hananel-Hazan mentioned this pull request Oct 29, 2026

Fix all known bugs for 0.3.5 (adopts #813 and #815) #831

Merged

6 tasks done

Hananel-Hazan closed this Oct 29, 2026

Navigation Menu

Repository navigation

Preserve the mean-field weight parameter during normalization - #813

Preserve the mean-field weight parameter during normalization#813
AHMETHAKANBEZIR1 wants to merge 3 commits into
BindsNET:masterfrom
AHMETHAKANBEZIR1:fix/meanfield-normalization-parameter

AHMETHAKANBEZIR1 commented Oct 21, 2026 •

edited

Loading

Uh oh!

Hananel-Hazan left a comment

Uh oh!

AHMETHAKANBEZIR1 commented Oct 23, 2026

Uh oh!

Hananel-Hazan commented Oct 29, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

AHMETHAKANBEZIR1 commented Oct 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Hananel-Hazan left a comment

Choose a reason for hiding this comment

Uh oh!

AHMETHAKANBEZIR1 commented Oct 23, 2026

Uh oh!

Hananel-Hazan commented Oct 29, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

AHMETHAKANBEZIR1 commented Oct 21, 2026 •

edited

Loading