Update pipelines (#326)

* add/cleaning n_class motor imagery pipelines * Try Configuration * DL benchmark, problem with the callbacks and yml file * DL benchmark without Callbacks * DL benchmark with Callbacks * DL benchmark with Callbacks * DL benchmark with Callbacks - Bug of test * [pre-commit.ci] auto fixes from pre-commit.com hooks * DL benchmark with Callbacks - Fix Suggestion * [pre-commit.ci] auto fixes from pre-commit.com hooks * Adding pipeline Augmented Covariance Matrix * [pre-commit.ci] auto fixes from pre-commit.com hooks * Fix Keras Name * Fix Keras Name and Add documentation Augmented Dataset * [pre-commit.ci] auto fixes from pre-commit.com hooks * Renaming the file to Keras EEGNet * Removing imported but not unused * Fixing B008 * Adding parameter --with deeplearning for CI * Restoring * Keras mapping with Callbacks and Optimizer. Added DL model: - ShallowConvNet - DeepConvNet EEGNet_8_2 * [pre-commit.ci] auto fixes from pre-commit.com hooks * Added DL model: - EEGITNet - EEGTCNet - EEGNeX * Added DL model: - EEGITNet - EEGTCNet - EEGNeX * [pre-commit.ci] auto fixes from pre-commit.com hooks * Tensforlow import bug * Tensforlow import bug * [pre-commit.ci] auto fixes from pre-commit.com hooks * Removing --no-dev flag. * Updating __init__ file to expose the deep learning methods. * Renaming the file, functions; Cleaning imports; Adding docs; snake_case style for variable; * Renaming the file, functions; Cleaning imports; Adding docs; snake_case style for variable; Trying to solve B008 (not define function in the args). * changing name for variable * Updating yaml with the new path * Fixing e-mail and renaming again. * Trying to solve the windows point. * tmp removing windows until I figure out how to work * adding the runs-on: ${{ matrix.os }} * Updating whats_new.rst * missed dot * updating models name * Reduced Epoch and Batch size for a fast example * Updated paradigm: - LeftRightImagery - MotorImagery Added Keras pipeline to Pipelines folder Added new pipelines: - CSP + SVM grid - EN Grid - FgMDM - regCSP+shLDA Grid - TSSVM Grid * [pre-commit.ci] auto fixes from pre-commit.com hooks * Modification requested by Sylavin * [pre-commit.ci] auto fixes from pre-commit.com hooks * Modification requested by Sylavin * Fix EEGTCNet. Add Standardization of input for: - EEGNet - ShallowNet - DeepConNet - EEGTCNet * [pre-commit.ci] auto fixes from pre-commit.com hooks * Add Standardization of input for every KERAS pipeline * trying to restore test on windows, without tensorflow * import tensorflow only when needed * print warning if tensorflow not installed * switch dev-dep to group docs * switch to docs and re-enable windows in test * correct error * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update moabb/pipelines/deep_learning.py Co-authored-by: Sylvain Chevallier <sylvain.chevallier@universite-paris-saclay.fr> * Update examples/plot_benchmark_DL.py * Apply suggestions from code review * fix precommit * correct example * correct typo * add names and references * add new class/fct in API * add setup_seed and set_download_dir in init and doc API * remove text --------- Co-authored-by: Sylvain Chevallier <sylain.chevallier@universite-paris-saclay.fr> Co-authored-by: CARRARA Igor <igor.carrara@inria.fr> Co-authored-by: Bru <a.bruno@aluno.ufabc.edu.br> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Igor Carrara <94047258+carraraig@users.noreply.github.com>
NeuroTechX · Mar 3, 2023 · f48b6c9 · f48b6c9
1 parent 9ec412f
commit f48b6c9
Show file tree

Hide file tree

Showing 43 changed files with 1,966 additions and 13 deletions.
diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml
@@ -44,10 +44,10 @@ jobs:
 
       - name: Install dependencies
         if: steps.cached-dataset-docs.outputs.cache-hit != 'true'
-        run: poetry install --no-interaction --no-root --with deeplearning
+        run: poetry install --no-interaction --no-root --with docs,deeplearning
 
       - name: Install library
-        run: poetry install --no-interaction --with deeplearning
+        run: poetry install --no-interaction --with docs,deeplearning
 
       - name: Build docs
         run: |

diff --git a/.github/workflows/test-devel.yml b/.github/workflows/test-devel.yml
@@ -46,10 +46,15 @@ jobs:
         if: |
           (runner.os != 'Windows') &&
           (steps.cached-poetry-dependencies.outputs.cache-hit != 'true')
-        run: poetry install --no-interaction --no-root --with deeplearning --no-dev
+        run: poetry install --no-interaction --no-root --with deeplearning
 
-      - name: Install library
-        run: poetry install --no-interaction --with deeplearning --no-dev
+      - name: Install library (Linux/OSX)
+        if: ${{ runner.os != 'Windows' }}
+        run: poetry install --no-interaction --with deeplearning
+
+      - name: Install library (Windows)
+        if: ${{ runner.os == 'Windows' }}
+        run: poetry install --no-interaction
 
       - name: Run tests
         run: |

diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
@@ -46,10 +46,16 @@ jobs:
         if: |
           (runner.os != 'Windows') &&
           (steps.cached-poetry-dependencies.outputs.cache-hit != 'true')
-        run: poetry install --no-interaction --no-root --with deeplearning --no-dev
+        run: poetry install --no-interaction --no-root --with deeplearning
+
+      - name: Install dependencies
+        if: |
+          (runner.os != 'Windows') &&
+          (steps.cached-poetry-dependencies.outputs.cache-hit != 'true')
+        run: poetry install --no-interaction --no-root --with deeplearning
 
       - name: Install library
-        run: poetry install --no-interaction  --with deeplearning --no-dev
+        run: poetry install --no-interaction  --with deeplearning
 
       - name: Run tests
         run: |

diff --git a/contexts/mi_n_class.yml b/contexts/mi_n_class.yml
@@ -0,0 +1,2 @@
+MotorImagery:
+  n_classes: 3
diff --git a/docs/source/pipelines.rst b/docs/source/pipelines.rst
@@ -17,8 +17,17 @@ Pipelines
     features.LogVariance
     features.FM
 	 features.ExtendedSSVEPSignal
+    features.AugmentedDataset
+    features.StandardScaler_Epoch
 	 csp.TRCSP
 	 classification.SSVEP_CCA
+    deep_learning.KerasDeepConvNet
+    deep_learning.KerasEEGITNet
+    deep_learning.KerasEEGNet_8_2
+    deep_learning.KerasEEGNeX
+    deep_learning.KerasEEGTCNet
+    deep_learning.KerasShallowConvNet
+
 
 ------------
 Base & Utils
@@ -30,3 +39,6 @@ Base & Utils
 
     utils.create_pipeline_from_config
     utils.FilterBank
+    utils_deep_model.EEGNet
+    utils_deep_model.EEGNet_TC
+    utils_deep_model.TCN_block
diff --git a/docs/source/utils.rst b/docs/source/utils.rst
@@ -25,3 +25,5 @@ Utils
     :template: function.rst
 
     set_log_level
+    setup_seed
+    set_download_dir
diff --git a/docs/source/whats_new.rst b/docs/source/whats_new.rst
@@ -22,6 +22,8 @@ Enhancements
 - Adding a comprehensive benchmarking function (:gh:`264` by `Divyesh Narayanan`_ and `Sylvain Chevallier`_)
 - Add meta-information for datasets in documentation (:gh:`317` by `Bruno Aristimunha`_)
 - Add GridSearchCV for different evaluation procedure (:gh:`319` by `Igor Carrara`_)
+- Add six deep learning models (Tensorflow), and build a tutorial to show to used the deep learning model (:gh:`326` by `Igor Carrara`_, `Bruno Aristimunha`_ and `Sylvain Chevallier`_)
+- Add a augmentation model to the pipeline (:gh:`326` by `Igor Carrara`_)
 
 Bugs
 ~~~~

diff --git a/examples/pipelines_DL/Keras_DeepConvNet.yml b/examples/pipelines_DL/Keras_DeepConvNet.yml
@@ -0,0 +1,40 @@
+name: Keras_DeepConvNet
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.1002/hbm.23730
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasDeepConvNet
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5
diff --git a/examples/pipelines_DL/Keras_EEGITNet.yml b/examples/pipelines_DL/Keras_EEGITNet.yml
@@ -0,0 +1,40 @@
+name: Keras_EEGITNet
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.1109/ACCESS.2022.3161489
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasEEGITNet
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5
diff --git a/examples/pipelines_DL/Keras_EEGNeX.yml b/examples/pipelines_DL/Keras_EEGNeX.yml
@@ -0,0 +1,40 @@
+name: Keras_EEGNeX
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.48550/arXiv.2207.12369
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasEEGNeX
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5
diff --git a/examples/pipelines_DL/Keras_EEGNet_8_2.yml b/examples/pipelines_DL/Keras_EEGNet_8_2.yml
@@ -0,0 +1,40 @@
+name: Keras_EEGNet_8_2
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.1088/1741-2552/aace8c
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasEEGNet_8_2
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5
diff --git a/examples/pipelines_DL/Keras_EEGTCNet.yml b/examples/pipelines_DL/Keras_EEGTCNet.yml
@@ -0,0 +1,40 @@
+name: Keras_EEGTCNet
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.1109/SMC42975.2020.9283028
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasEEGTCNet
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5
diff --git a/examples/pipelines_DL/Keras_ShallowConvNet.yml b/examples/pipelines_DL/Keras_ShallowConvNet.yml
@@ -0,0 +1,40 @@
+name: Keras_ShallowConvNet
+
+paradigms:
+  - LeftRightImagery
+  - MotorImagery
+
+citations:
+  - https://doi.org/10.1002/hbm.23730
+
+pipeline:
+  - name: StandardScaler_Epoch
+    from: moabb.pipelines.features
+
+  - name: KerasShallowConvNet
+    from: moabb.pipelines.deep_learning
+    parameters:
+      loss: "sparse_categorical_crossentropy"
+      optimizer:
+        - name: Adam
+          from: tensorflow.keras.optimizers.legacy
+          parameters:
+            learning_rate: 0.0009
+      epochs: 2
+      batch_size: 1
+      verbose: 0
+      random_state: 42
+      validation_split: 0.2
+      callbacks:
+        - name: EarlyStopping
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+
+        - name: ReduceLROnPlateau
+          from: tensorflow.keras.callbacks
+          parameters:
+            monitor: "val_loss"
+            patience: 300
+            factor: 0.5