MLP

This is an unofficial PyTorch implementation created by Ignacio Oguiza (oguiza@timeseriesAI.co) based on:

Fawaz, H. I., Forestier, G., Weber, J., Idoumghar, L., & Muller, P. A. (2019). Deep learning for time series classification: a review. Data Mining and Knowledge Discovery, 33(4), 917-963.

Official MLP TensorFlow implementation: https://github.com/hfawaz/dl-4-tsc/blob/master/classifiers/mlp.py

---

source

MLP

 MLP (c_in, c_out, seq_len, layers=[500, 500, 500], ps=[0.1, 0.2, 0.2],
      act=ReLU(inplace=True), use_bn=False, bn_final=False,
      lin_first=False, fc_dropout=0.0, y_range=None)

Same as nn.Module, but no need for subclasses to call super().__init__

bs = 16
nvars = 3
seq_len = 128
c_out = 2
xb = torch.rand(bs, nvars, seq_len)
model = MLP(nvars, c_out, seq_len)
test_eq(model(xb).shape, (bs, c_out))
model

MLP(
  (flatten): Reshape(bs)
  (mlp): ModuleList(
    (0): LinBnDrop(
      (0): Dropout(p=0.1, inplace=False)
      (1): Linear(in_features=384, out_features=500, bias=True)
      (2): ReLU(inplace=True)
    )
    (1): LinBnDrop(
      (0): Dropout(p=0.2, inplace=False)
      (1): Linear(in_features=500, out_features=500, bias=True)
      (2): ReLU(inplace=True)
    )
    (2): LinBnDrop(
      (0): Dropout(p=0.2, inplace=False)
      (1): Linear(in_features=500, out_features=500, bias=True)
      (2): ReLU(inplace=True)
    )
  )
  (head): Sequential(
    (0): LinBnDrop(
      (0): Linear(in_features=500, out_features=2, bias=True)
    )
  )
)