merge with master #722

Hananel-Hazan · 2025-11-07T01:12:21Z

No description provided.

docs: Update README badges - Collaboration Network

--- updated-dependencies: - dependency-name: jupyterlab dependency-version: 4.4.8 dependency-type: direct:development ... Signed-off-by: dependabot[bot] <[email protected]>

Bump jupyterlab from 4.4.3 to 4.4.8

Bumps [pillow](https://github.com/python-pillow/Pillow) from 11.2.1 to 11.3.0. - [Release notes](https://github.com/python-pillow/Pillow/releases) - [Changelog](https://github.com/python-pillow/Pillow/blob/main/CHANGES.rst) - [Commits](python-pillow/Pillow@11.2.1...11.3.0) --- updated-dependencies: - dependency-name: pillow dependency-version: 11.3.0 dependency-type: indirect ... Signed-off-by: dependabot[bot] <[email protected]>

Bump pillow from 11.2.1 to 11.3.0

…ling across multiple files

…ling and warnings

…snet into lowering_precision

…sion and batch_eth_mnist scripts

SparseConnection support

…structor calls across multiple classes.

Lowering precision

Copilot

Pull Request Overview

This pull request implements sparse tensor support for BindsNET, enabling memory-efficient representation of neural network connections with low sparsity. The changes introduce sparse tensor capabilities across connections, features, monitors, and learning rules, along with benchmarking tools and documentation to demonstrate performance characteristics.

Key Changes

Added sparse tensor support to network connections, topology features, monitors, and learning rules
Introduced w_dtype parameter for weight precision control across connection types
Added benchmark scripts to compare sparse vs dense tensor performance and different precision levels
Updated DiehlAndCook2015 model to use MulticompartmentConnection with sparse support

Reviewed Changes

Copilot reviewed 13 out of 88 changed files in this pull request and generated 11 comments.

Show a summary per file

File	Description
test/network/test_connections.py	Added test coverage for `SparseConnection` with wmin/wmax validation
poetry.lock	Updated Poetry and package dependencies (JupyterLab, Pillow, pandas, etc.)
examples/mnist/batch_eth_mnist.py	Added sparse tensor and precision control support to MNIST example
examples/benchmark/sparse_vs_dense_tensors.py	New benchmark comparing memory usage and runtime of sparse vs dense tensors
examples/benchmark/lowering_precision.py	New benchmark measuring impact of reduced precision (float16 vs float32)
docs/source/guide/guide_part_i.rst	Added comprehensive documentation for sparse tensors and precision control
bindsnet/network/topology_features.py	Implemented sparse tensor support in feature classes with dtype control
bindsnet/network/topology.py	Added w_dtype parameter to all connection types and simplified SparseConnection
bindsnet/network/monitors.py	Added sparse parameter to enable sparse tensor recording in monitors
bindsnet/models/models.py	Updated DiehlAndCook2015 to use MulticompartmentConnection with sparse support
bindsnet/learning/learning.py	Added sparse tensor handling in learning rule updates
bindsnet/learning/MCC_learning.py	Implemented sparse tensor support in MCC learning rules
bindsnet/evaluation/evaluation.py	Added sparse tensor conversion in evaluation functions
README.md	Updated badges to replace deprecated/broken links
.gitignore	Added recursive pyc file pattern

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2025-11-07T01:16:53Z

bindsnet/learning/MCC_learning.py

+                if self.feature_value.is_sparse:
+                    self.feature_value -= (
+                        torch.bmm(source_s, target_x) * self.connection.dt
+                    ).to_sparse()


When self.feature_value.is_sparse is True, the reduction function is skipped but the batch dimension is not reduced. This may result in incorrect tensor shapes. The update should reduce the batch dimension using self.reduction before converting to sparse: self.feature_value -= self.reduction(torch.bmm(source_s, target_x), dim=0).to_sparse() * self.connection.dt.

Copilot · 2025-11-07T01:16:53Z

examples/mnist/batch_eth_mnist.py

        if plot:
            image = batch["image"][:, 0].view(28, 28)
            inpt = inputs["X"][:, 0].view(time, 784).sum(0).view(28, 28)
            lable = batch["label"][0]


Typo in variable name: 'lable' should be 'label'.

@copilot open a new pull request to apply changes based on this feedback

Copilot · 2025-11-07T01:16:53Z

bindsnet/network/topology_features.py

+            warnings.warn(
+                f"Provided value has data type {value.dtype} but parameter w_dtype is {value_dtype}"


The warning message refers to "parameter w_dtype" but the actual parameter name is "value_dtype". The message should say "parameter value_dtype is" instead of "parameter w_dtype is".

@copilot open a new pull request to apply changes based on this feedback

Copilot · 2025-11-07T01:16:53Z

bindsnet/learning/learning.py

        ) and not isinstance(self, NoOp):
-            self.connection.w.clamp_(self.connection.wmin, self.connection.wmax)
+            if self.connection.w.is_sparse:
+                raise Exception("SparseConnection isn't supported for wmin\\wmax")


The exception message contains an incorrect escape sequence. The backslash should be properly escaped or the message should use a forward slash. It should be "wmin/wmax" or "wmin\\wmax" instead of "wmin\wmax".

@copilot open a new pull request to apply changes based on this feedback

Copilot · 2025-11-07T01:16:54Z

bindsnet/evaluation/evaluation.py


            # Compute average firing rates for this label.
+            selected_spikes = torch.index_select(
+                spikes, dim=0, index=torch.tensor(indices)


The torch.tensor(indices) call is unnecessary and may cause performance issues or device mismatches. The indices variable is already a tensor on the same device as spikes. It should be used directly: torch.index_select(spikes, dim=0, index=indices).

Suggested change

spikes, dim=0, index=torch.tensor(indices)

spikes, dim=0, index=indices

Copilot · 2025-11-07T01:16:54Z

bindsnet/learning/MCC_learning.py

+                        torch.bmm(source_x, target_s) * self.connection.dt
+                    ).to_sparse()


When self.feature_value.is_sparse is True, the reduction function is skipped but the batch dimension is not reduced. This may result in incorrect tensor shapes. The update should reduce the batch dimension using self.reduction before converting to sparse: self.feature_value += self.reduction(torch.bmm(source_x, target_s), dim=0).to_sparse() * self.connection.dt.

Suggested change

torch.bmm(source_x, target_s) * self.connection.dt

).to_sparse()

self.reduction(torch.bmm(source_x, target_s), dim=0).to_sparse() * self.connection.dt

)

Copilot · 2025-11-07T01:16:54Z

test/network/test_connections.py

                    ):
                        return

+                    # SparseConnection isn't supported for wmin\\wmax


The comment contains an incorrect escape sequence. The backslash should be properly escaped or the comment should use a forward slash. It should be "wmin/wmax" or "wmin\\wmax" instead of "wmin\wmax".

Copilot · 2025-11-07T01:16:55Z

examples/benchmark/sparse_vs_dense_tensors.py

+        "======================= ====================== ====================== ===================="
+    )
+    print(
+        "Sparse (megabytes used) Dense (megabytes used) Ratio (Sparse/Dense) % % of non zero values"


The table header has "% % of non zero values" with duplicated percent signs. This should likely be just "% of non zero values".

@copilot open a new pull request to apply changes based on this feedback

Copilot · 2025-11-07T01:16:55Z

docs/source/guide/guide_part_i.rst

+======================= ====================== ====================== ====================
+Sparse (megabytes used) Dense (megabytes used) Ratio (Sparse/Dense) % % of non zero values
+======================= ====================== ====================== ====================


The table header has "% % of non zero values" with duplicated percent signs. This should likely be just "% of non zero values".

@copilot open a new pull request to apply changes based on this feedback

Copilot · 2025-11-07T01:16:55Z

bindsnet/models/models.py

 import torch
 from scipy.spatial.distance import euclidean
 from torch.nn.modules.utils import _pair
+from torch import device


Import of 'device' is not used.

@copilot open a new pull request to apply changes based on this feedback

Hananel-Hazan · 2025-11-07T01:20:50Z

@copilot open a new pull request to apply changes based on the comments in this thread

n-shevko and others added 30 commits January 23, 2025 14:21

SparseConnection support

0308b27

Test for SparseConnection

d1d3e42

Sparsity for MulticompartmentConnection

471d455

Sparse batch_eth_mnist

8b1cb4e

Sparsity for s monitor

12d8e03

Add batch dimension in case if sparse == True

2c0760d

Add doc for sparse=True

9678254

Make --sparse influence Monitors

fd0ef9b

Note about performance

ec63566

Lowering precision for old connections

e0f9bf2

w_dtype for example

e9dfcbf

Make device and batch_size optional

bed0623

Add sparse support for other learning rules

326d270

Sparse tensors for monitors explanation

95626a9

Add runtime experiment

9ff6f01

Use cuda

3c9528a

Make tensor smaller to fit in 7gb of gpu memory

717ee17

Documentation update

988cb91

Merge pull request #711 from BindsNET/hananel

88f01a8

docs: Update README badges

85c8cb1

docs: Update neuromorphic badge in README

6e938f4

Merge pull request #712 from neural-loop/collaboration-network

5833f9d

docs: Update README badges - Collaboration Network

Update badges README.md

80ba5e5

Bump jupyterlab from 4.4.3 to 4.4.8

a7dc702

--- updated-dependencies: - dependency-name: jupyterlab dependency-version: 4.4.8 dependency-type: direct:development ... Signed-off-by: dependabot[bot] <[email protected]>

Merge pull request #720 from BindsNET/dependabot/pip/jupyterlab-4.4.8

1c8cb82

Bump jupyterlab from 4.4.3 to 4.4.8

Merge pull request #721 from BindsNET/dependabot/pip/pillow-11.3.0

731206e

Bump pillow from 11.2.1 to 11.3.0

Refactor code for improved readability and consistency in tensor hand…

6b6e4b5

…ling across multiple files

Refactor code for improved readability and consistency in weight hand…

08f4bd0

…ling and warnings

Docs

5aceb2e

n-shevko and others added 6 commits November 6, 2025 17:25

Merge branch 'lowering_precision' of https://github.com/BindsNET/bind…

2adfcba

…snet into lowering_precision

Refactor string formatting for improved readability in lowering_preci…

b5c608b

…sion and batch_eth_mnist scripts

Merge pull request #703 from BindsNET/sparse_connection

4c894b7

SparseConnection support

Merge branch 'master' into lowering_precision

f5a716d

Refactor code for improved readability by formatting arguments in con…

f19692c

…structor calls across multiple classes.

Merge pull request #707 from BindsNET/lowering_precision

d072219

Lowering precision

Hananel-Hazan requested a review from Copilot November 7, 2025 01:12

Merge branch 'hananel' into master

fd3b904

Hananel-Hazan merged commit a2840ad into hananel Nov 7, 2025
4 of 5 checks passed

Copilot AI reviewed Nov 7, 2025

View reviewed changes

		warnings.warn(
		f"Provided value has data type {value.dtype} but parameter w_dtype is {value_dtype}"

	spikes, dim=0, index=torch.tensor(indices)
	spikes, dim=0, index=indices

		torch.bmm(source_x, target_s) * self.connection.dt
		).to_sparse()

merge with master #722

merge with master #722

Uh oh!

Conversation

Hananel-Hazan commented Nov 7, 2025

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Key Changes

Reviewed Changes

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

Hananel-Hazan commented Nov 7, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants