AdaptiveMotorControlLab · MMathisLab · Apr 17, 2025 · Apr 14, 2025 · Apr 14, 2025 · Apr 15, 2025
diff --git a/AUTHORS.md b/AUTHORS.md
@@ -2,27 +2,25 @@
 
 
 
-CEBRA was initially developed by **Mackenzie Mathis** and **Steffen Schneider** (2021+), who are co-inventors on the patent application [WO2023143843](https://infoscience.epfl.ch/entities/patent/0d9debed-4d22-47b7-bad1-f211e7010323). 
-**Jin Hwa Lee** contributed significantly to our first paper:  
+CEBRA was initially developed by **Mackenzie Mathis** and **Steffen Schneider** (2021+), who are co-inventors on the patent application [WO2023143843](https://infoscience.epfl.ch/entities/patent/0d9debed-4d22-47b7-bad1-f211e7010323).
+**Jin Hwa Lee** contributed significantly to our first paper:
 
-> **Schneider, S., Lee, J.H., & Mathis, M.W.**  
-> [*Learnable latent embeddings for joint behavioural and neural analysis.*](https://doi.org/10.1038/s41586-023-06031-6) 
+> **Schneider, S., Lee, J.H., & Mathis, M.W.**
+> [*Learnable latent embeddings for joint behavioural and neural analysis.*](https://doi.org/10.1038/s41586-023-06031-6)
 > Nature 617, 360–368 (2023)
 
-CEBRA is actively developed by [**Mackenzie Mathis**](https://www.mackenziemathislab.org/) and [**Steffen Schneider**](https://dynamical-inference.ai/) and their labs.  
+CEBRA is actively developed by [**Mackenzie Mathis**](https://www.mackenziemathislab.org/) and [**Steffen Schneider**](https://dynamical-inference.ai/) and their labs.
 
-It is a publicly available tool that has benefited from contributions and suggestions from many individuals: [CEBRA/graphs/contributors](https://github.com/AdaptiveMotorControlLab/CEBRA/graphs/contributors).  
+It is a publicly available tool that has benefited from contributions and suggestions from many individuals: [CEBRA/graphs/contributors](https://github.com/AdaptiveMotorControlLab/CEBRA/graphs/contributors).
 
-## CEBRA Extensions 
+## CEBRA Extensions
 
-### 2023  
-- **Steffen Schneider, Rodrigo González Laiz, Markus Frey, Mackenzie W. Mathis**  
-  [*Identifiable attribution maps using regularized contrastive learning.*](https://sslneurips23.github.io/paper_pdfs/paper_80.pdf) 
+### 2023
+- **Steffen Schneider, Rodrigo González Laiz, Markus Frey, Mackenzie W. Mathis**
+  [*Identifiable attribution maps using regularized contrastive learning.*](https://sslneurips23.github.io/paper_pdfs/paper_80.pdf)
   NeurIPS 4th Workshop on Self-Supervised Learning: Theory and Practice (2023)
 
-### 2025  
-- **Steffen Schneider, Rodrigo González Laiz, Anastasiia Filippova, Markus Frey, Mackenzie W. Mathis**  
-  [*Time-series attribution maps with regularized contrastive learning.*](https://openreview.net/forum?id=aGrCXoTB4P)  
+### 2025
+- **Steffen Schneider, Rodrigo González Laiz, Anastasiia Filippova, Markus Frey, Mackenzie W. Mathis**
+  [*Time-series attribution maps with regularized contrastive learning.*](https://openreview.net/forum?id=aGrCXoTB4P)
   AISTATS (2025)
-
-
diff --git a/Dockerfile b/Dockerfile
@@ -40,7 +40,7 @@ RUN make dist
 FROM cebra-base
 
 # install the cebra wheel
-ENV WHEEL=cebra-0.5.0rc1-py3-none-any.whl
+ENV WHEEL=cebra-0.5.0-py3-none-any.whl
 WORKDIR /build
 COPY --from=wheel /build/dist/${WHEEL} .
 RUN pip install --no-cache-dir ${WHEEL}'[dev,integrations,datasets]'

diff --git a/Makefile b/Makefile
@@ -1,4 +1,4 @@
-CEBRA_VERSION := 0.5.0rc1
+CEBRA_VERSION := 0.5.0
 
 dist:
 	python3 -m pip install virtualenv

diff --git a/PKGBUILD b/PKGBUILD
@@ -1,7 +1,7 @@
 # Maintainer: Steffen Schneider <stes@hey.com>
 pkgname=python-cebra
 _pkgname=cebra
-pkgver=0.5.0rc1
+pkgver=0.5.0
 pkgrel=1
 pkgdesc="Consistent Embeddings of high-dimensional Recordings using Auxiliary variables"
 url="https://cebra.ai"

diff --git a/cebra/__init__.py b/cebra/__init__.py
@@ -66,7 +66,7 @@
 
 import cebra.integrations.sklearn as sklearn
 
-__version__ = "0.5.0rc1"
+__version__ = "0.5.0"
 __all__ = ["CEBRA"]
 __allow_lazy_imports = False
 __lazy_imports = {}

diff --git a/cebra/integrations/plotly.py b/cebra/integrations/plotly.py
@@ -27,8 +27,8 @@
 import numpy as np
 import numpy.typing as npt
 import plotly.graph_objects
-import torch
 import plotly.graph_objects as go
+import torch
 
 from cebra.integrations.matplotlib import _EmbeddingPlot
 
@@ -153,17 +153,18 @@ def _plot_3d(self, **kwargs) -> plotly.graph_objects.Figure:
 
 
 def plot_embedding_interactive(
-    embedding: Union[npt.NDArray, torch.Tensor],
-    embedding_labels: Optional[Union[npt.NDArray, torch.Tensor, str]] = "grey",
-    axis: Optional["go.Figure"] = None,
-    markersize: float = 1,
-    idx_order: Optional[Tuple[int]] = None,
-    alpha: float = 0.4,
-    cmap: str = "cool",
-    title: str = "Embedding",
-    figsize: Tuple[int] = (5, 5),
-    dpi: int = 100,
-    **kwargs,
+        embedding: Union[npt.NDArray, torch.Tensor],
+        embedding_labels: Optional[Union[npt.NDArray, torch.Tensor,
+                                         str]] = "grey",
+        axis: Optional["go.Figure"] = None,
+        markersize: float = 1,
+        idx_order: Optional[Tuple[int]] = None,
+        alpha: float = 0.4,
+        cmap: str = "cool",
+        title: str = "Embedding",
+        figsize: Tuple[int] = (5, 5),
+        dpi: int = 100,
+        **kwargs,
 ) -> "go.Figure":
     """Plot embedding in a 3D dimensional space.
 

diff --git a/docs/root/index.html b/docs/root/index.html
@@ -145,16 +145,16 @@ <h4>
                 </div>
 
                 <div class="row mb-5 mt-4">
-                    <p>CEBRA is a machine-learning method that can be used to 
+                    <p>CEBRA is a machine-learning method that can be used to
                         compress time series in a way that reveals otherwise hidden
-                         structures in the variability of the data. It excels on 
-                         behavioural and neural data recorded simultaneously. 
+                         structures in the variability of the data. It excels on
+                         behavioural and neural data recorded simultaneously.
                          We have shown it can be used to decode the activity from the
                           visual cortex of the mouse brain to reconstruct a viewed video,
                            to decode trajectories from the sensoirmotor cortex of primates,
                             and for decoding position during navigation. For these use cases
                             and other demos see our <a href="https://cebra.ai/docs/" style="color: #6235E0;">Documentation</a>.</p>
-                
+
                 </div>
 
                 <div class="row">
@@ -171,12 +171,12 @@ <h3><i class="fas fa-play-circle"></i> Demo Applications</h3>
                     <div class="col-md-6 mb-2">
                         <!-- Embedding the Plotly figure using iframe -->
                         <div style="position: relative; height: 315px; overflow: hidden; margin-bottom: 1rem;">
-                            <iframe src="static/img/hippocampus_posdir3_full.html" 
+                            <iframe src="static/img/hippocampus_posdir3_full.html"
                                     style="position: absolute; top: -140px; left: -5%; width: 110%; height: 150%; border: none; transform: scale(0.85); transform-origin: top center;"
                                     scrolling="no">
                             </iframe>
                         </div>
-                        
+
                         <p style="margin-top: -70px;">Interactive visualization of the CEBRA embedding for the rat hippocampus data. This 3D plot shows how neural activity is mapped to a lower-dimensional space that correlates with the animal's position and movement direction. <a href="https://colab.research.google.com/github/AdaptiveMotorControlLab/CEBRA-demos/blob/main/Demo_hippocampus.ipynb" target="_blank" style="color: #6235E0;"><i class="fas fa-external-link-alt"></i> Open In Colaboratory</a></p>
                     </div>
                 </div>
@@ -191,7 +191,7 @@ <h3><i class="fas fa-play-circle"></i> Demo Applications</h3>
                         <p>CEBRA applied to mouse primary visual cortex, collected at the Allen Institute (de Vries et al. 2020, Siegle et al. 2021). 2-photon and Neuropixels recordings are embedded with CEBRA using DINO frame features as labels.
                         The embedding is used to decode the video frames using a kNN decoder on the CEBRA-Behavior embedding from the test set.</p>
                     </div>
-                    
+
                     <div class="col-md-6 mb-2">
                         <!-- YouTube embed for CEBRA on M1 and S1 neural data with cleaner styling -->
                         <video width="100%" autoplay loop muted preload="auto">
@@ -205,7 +205,7 @@ <h3><i class="fas fa-play-circle"></i> Demo Applications</h3>
 
                 <div class="row mt-4">
                     <h3><i class="fas fa-newspaper"></i> Publications</h3>
-                    
+
                     <div class="col-12">
                         <div class="paper-card">
                             <div class="paper-title">Learnable latent embeddings for joint behavioural and neural analysis</div>
@@ -215,7 +215,7 @@ <h3><i class="fas fa-newspaper"></i> Publications</h3>
                             <a href="https://arxiv.org/abs/2204.00673" target="_blank" class="btn btn-link" style="color: #6235E0;"><i class="fas fa-file-alt"></i> Preprint</a>
                         </div>
                     </div>
-                    
+
                     <div class="col-12">
                         <div class="paper-card">
                             <div class="paper-title">Time-series attribution maps with regularized contrastive learning</div>
@@ -230,7 +230,7 @@ <h3><i class="fas fa-newspaper"></i> Publications</h3>
 
                 <div class="row mt-4">
                     <h3><i class="fas fa-certificate"></i> Patent Information</h3>
-                    
+
                     <div class="col-12">
                         <div class="paper-card">
                             <div class="paper-title">Patent Pending</div>
@@ -299,7 +299,7 @@ <h3>
                         </small>
                     </div>
                 </div>
-                
+
                 <div class="row justify-content-md-center">
                     <div class="col-sm-10 rounded p-3 m-2" style="background-color: rgb(20,20,20);">
                         <small class="code">
@@ -325,13 +325,13 @@ <h3>
                     <p>
                         CEBRA has been cited in numerous high-impact publications across neuroscience, machine learning, and related fields. Our work has influenced research in neural decoding, brain-computer interfaces, computational neuroscience, and machine learning methods for time-series analysis.
                     </p>
-                    
+
                     <div class="col-12 text-center mb-4">
                         <a href="https://scholar.google.com/scholar?oi=bibs&hl=en&cites=5385393104765622341&as_sdt=5" target="_blank" class="btn btn-outline-light btn-lg">
                             <i class="fas fa-graduation-cap"></i> View All Citations on Google Scholar
                         </a>
                     </div>
-                    
+
                     <div class="col-12">
                         <div class="paper-card">
                             <p class="mb-0">Our research has been cited in proceedings and journals including <span class="badge bg-light text-dark">Nature</span> <span class="badge bg-light text-dark">Science</span> <span class="badge bg-light text-dark">ICML</span> <span class="badge bg-light text-dark">Nature Neuroscience</span> <span class="badge bg-light text-dark">ICML</span> <span class="badge bg-light text-dark">Neuron</span> <span class="badge bg-light text-dark">NeurIPS</span> <span class="badge bg-light text-dark">ICLR</span> and others.</p>
@@ -342,8 +342,8 @@ <h3>
                 <div class="row justify-content-center mt-5 mb-3">
                     <div class="col-md-12 text-center">
                         <a href="https://www.epfl.ch/" target="_blank">
-                            <img src="https://images.squarespace-cdn.com/content/v1/57f6d51c9f74566f55ecf271/00b1fa45-9246-4914-86ee-4a01bb3fb60b/logo.png?format=2500w" 
-                                alt="MLAI Logo" 
+                            <img src="https://images.squarespace-cdn.com/content/v1/57f6d51c9f74566f55ecf271/00b1fa45-9246-4914-86ee-4a01bb3fb60b/logo.png?format=2500w"
+                                alt="MLAI Logo"
                                 style="max-width: 600px;">
                         </a>
                         <div class="mt-3">

diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -81,7 +81,6 @@ def get_years(start_year=2021):
     "sphinx_gallery.load_style",
 ]
 
-
 coverage_show_missing_items = True
 panels_add_bootstrap_css = False
 
@@ -147,7 +146,8 @@ def get_years(start_year=2021):
 html_context = {
     "default_mode": "light",
     "switcher": {
-        "version_match": "latest",  # Adjust this dynamically per version
+        "version_match":
+            "latest",  # Adjust this dynamically per version
         "versions": [
             ("latest", "/latest/"),
             ("v0.2.0", "/v0.2.0/"),
@@ -156,7 +156,8 @@ def get_years(start_year=2021):
             ("v0.5.0rc1", "/v0.5.0rc1/"),
         ],
     },
-    "navbar_start": ["version-switcher", "navbar-logo"],  # Place the dropdown above the logo
+    "navbar_start": ["version-switcher",
+                     "navbar-logo"],  # Place the dropdown above the logo
 }
 
 # More info on theme options:
@@ -220,7 +221,7 @@ def get_years(start_year=2021):
 
 nbsphinx_thumbnails = {
     "demo_notebooks/CEBRA_best_practices":
-    "_static/thumbnails/cebra-best.png",
+        "_static/thumbnails/cebra-best.png",
     "demo_notebooks/Demo_primate_reaching":
         "_static/thumbnails/ForelimbS1.png",
     "demo_notebooks/Demo_hippocampus":

diff --git a/docs/source/usage.rst b/docs/source/usage.rst
@@ -1218,36 +1218,36 @@ Putting all previous snippet examples together, we obtain the following pipeline
         output_dimension = 8,
         verbose = False
      )
-    
+
      # 2. Load example data
      neural_data = cebra.load_data(file="neural_data.npz", key="neural")
      new_neural_data = cebra.load_data(file="neural_data.npz", key="new_neural")
      continuous_label = cebra.load_data(file="auxiliary_behavior_data.h5", key="auxiliary_variables", columns=["continuous1", "continuous2", "continuous3"])
      discrete_label = cebra.load_data(file="auxiliary_behavior_data.h5", key="auxiliary_variables", columns=["discrete"]).flatten()
-    
-    
+
+
      assert neural_data.shape == (100, 3)
      assert new_neural_data.shape == (100, 4)
      assert discrete_label.shape == (100, )
      assert continuous_label.shape == (100, 3)
-    
+
      # 3. Split data and labels into train/validation
      from sklearn.model_selection import train_test_split
-    
+
      split_idx = int(0.8 * len(neural_data))
      # suggestion: 5%-20% depending on your dataset size; note that this splits the
      # into an early and late part, which might not be ideal for your data/experiment!
      # As a more involved alternative, consider e.g. a nested time-series split.
-    
+
      train_data = neural_data[:split_idx]
      valid_data = neural_data[split_idx:]
-    
+
      train_continuous_label = continuous_label[:split_idx]
      valid_continuous_label = continuous_label[split_idx:]
-    
+
      train_discrete_label = discrete_label[:split_idx]
      valid_discrete_label = discrete_label[split_idx:]
-    
+
      # 4. Fit the model
      # time contrastive learning
      cebra_model.fit(train_data)
@@ -1257,29 +1257,29 @@ Putting all previous snippet examples together, we obtain the following pipeline
      cebra_model.fit(train_data, train_continuous_label)
      # mixed behavior contrastive learning
      cebra_model.fit(train_data, train_discrete_label, train_continuous_label)
-    
-    
+
+
      # 5. Save the model
      tmp_file = Path(tempfile.gettempdir(), 'cebra.pt')
      cebra_model.save(tmp_file)
-    
+
      # 6. Load the model and compute an embedding
      cebra_model = cebra.CEBRA.load(tmp_file)
      train_embedding = cebra_model.transform(train_data)
      valid_embedding = cebra_model.transform(valid_data)
-    
+
      assert train_embedding.shape == (80, 8) # TODO(user): change to split ratio & output dim
      assert valid_embedding.shape == (20, 8) # TODO(user): change to split ratio & output dim
-    
+
      # 7. Evaluate the model performance (you can also check the train_data)
      goodness_of_fit = cebra.sklearn.metrics.goodness_of_fit_score(cebra_model,
                                                           valid_data,
                                                           valid_discrete_label,
                                                           valid_continuous_label)
-    
+
      # 8. Adapt the model to a new session
      cebra_model.fit(new_neural_data, adapt = True)
-    
+
      # 9. Decode discrete labels behavior from the embedding
      decoder = cebra.KNNDecoder()
      decoder.fit(train_embedding, train_discrete_label)

diff --git a/reinstall.sh b/reinstall.sh
@@ -15,7 +15,7 @@ pip uninstall -y cebra
 # Get version info after uninstalling --- this will automatically get the
 # most recent version based on the source code in the current directory.
 # $(tools/get_cebra_version.sh)
-VERSION=0.5.0rc1
+VERSION=0.5.0
 echo "Upgrading to CEBRA v${VERSION}"
 
 # Upgrade the build system (PEP517/518 compatible)

diff --git a/setup.cfg b/setup.cfg
@@ -39,7 +39,9 @@ install_requires =
     scipy
     torch>=2.4.0
     tqdm
-    matplotlib
+    # NOTE(stes): Remove pin once https://github.com/AdaptiveMotorControlLab/CEBRA/issues/240
+    # is resolved.
+    matplotlib<3.11
     requests
 
 [options.extras_require]