Tested model:
=============
------- notif_encoder:
NotifEncoder(
  (model): Sequential(
    (0): Linear(in_features=29, out_features=64, bias=True)
    (1): ReLU()
    (2): Linear(in_features=64, out_features=32, bias=True)
    (3): Dropout(p=0.2, inplace=False)
  )
)

------- evnt_encoder:
EvntEncoder(
  (model): Sequential(
    (0): Linear(in_features=11, out_features=64, bias=True)
    (1): ReLU()
    (2): Linear(in_features=64, out_features=32, bias=True)
    (3): Dropout(p=0.2, inplace=False)
  )
)

------- info_encoder:
InfoEncoder(
  (model): Sequential(
    (0): Linear(in_features=28, out_features=64, bias=True)
    (1): PReLU(num_parameters=1)
    (2): Linear(in_features=64, out_features=32, bias=True)
    (3): Dropout(p=0.2, inplace=False)
  )
)

------- embedder:
ConvEmbedderSimple(
  (adapter): Linear(in_features=32, out_features=32, bias=True)
  (representer): ConvRepresenter(
    (convs): ParameterList(
        (0): Object of type: ConvRepresenterBlock
        (1): Object of type: ConvRepresenterBlock
      (0): ConvRepresenterBlock(
        (model): Sequential(
          (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=(1,))
          (1): PReLU(num_parameters=1)
          (2): Conv1d(64, 32, kernel_size=(3,), stride=(1,), padding=(1,))
          (3): PReLU(num_parameters=1)
        )
      )
      (1): ConvRepresenterBlock(
        (model): Sequential(
          (0): Conv1d(32, 64, kernel_size=(3,), stride=(1,), padding=(1,))
          (1): PReLU(num_parameters=1)
          (2): Conv1d(64, 32, kernel_size=(3,), stride=(1,), padding=(1,))
          (3): PReLU(num_parameters=1)
        )
      )
    )
  )
  (final): Conv1d(32, 32, kernel_size=(22,), stride=(1,))
)

(Training dataset class occurences: {4: 45, 3: 176, 1: 46, 2: 1}.)
svm finished: True; SV cnt 109;
Trained SVM with LINEAR kernel with loss 920000.0000000066 and 109 outliers.
svm finished: True; SV cnt 97;
Trained SVM with POLYNOMIAL (deg 2) kernel with loss 920000.0000057232 and 97 outliers.
svm finished: True; SV cnt 99;
Trained SVM with GAUSSIAN (sigma 1) kernel with loss 920000.0000057232 and 97 outliers.
Results of classification of specially annotated notifications:
  IDs of clssified notificaions: [11168292, 11169523, 11173814, 9961382, 9962202, 9971171, 10168557, 10213020, 10275769]
  classes of classified notifications: [3, 3, 1, 3, 3, 1, 3, 2, 3]
  SVM with LINEAR kernel: tensor([1., 1., 1., 1., 1., 1., 1., 1., 1.], dtype=torch.float64)
  SVM with POLYNOMIAL kernel: tensor([1., 1., 1., 1., 1., 1., 1., 1., 1.], dtype=torch.float64)
  SVM with GAUSSIAN kernel: tensor([1., 1., 1., 1., 1., 1., 1., 1., 1.], dtype=torch.float64)
  logistic regression: tensor([[9.9913e-01, 8.7022e-04],
        [9.9989e-01, 1.1063e-04],
        [5.8318e-01, 4.1682e-01],
        [9.8751e-01, 1.2493e-02],
        [9.8005e-01, 1.9947e-02],
        [7.5498e-01, 2.4502e-01],
        [1.0000e+00, 4.7845e-07],
        [9.9959e-01, 4.1077e-04],
        [9.7019e-01, 2.9814e-02]], grad_fn=<SoftmaxBackward0>)
  MLP classifier: tensor([[1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.],
        [1., 0.]], grad_fn=<SoftmaxBackward0>)
Results on test set:
  SVM with LINEAR kernel: 1.0 (13/13)
  SVM with POLYNOMIAL kernel: 1.0 (13/13)
  SVM with GAUSSIAN kernel: 1.0 (13/13)
  logistic regression: 0.8461538553237915 (11/13)
  MLP classifier: 1.0 (13/13)
  (class occurence counts: {3: 11, 4: 2})