OmniScaleCNN

This is an unofficial PyTorch implementation created by Ignacio Oguiza - oguiza@timeseriesAI.co

---

source

generate_layer_parameter_list

 generate_layer_parameter_list (start, end, layers, in_channel=1)

---

source

get_out_channel_number

 get_out_channel_number (paramenter_layer, in_channel, prime_list)

---

source

get_Prime_number_in_a_range

 get_Prime_number_in_a_range (start, end)

---

source

OmniScaleCNN

 OmniScaleCNN (c_in, c_out, seq_len, layers=[1024, 229376],
               few_shot=False)

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

---

source

build_layer_with_layer_parameter

 build_layer_with_layer_parameter (layer_parameters)

formerly build_layer_with_layer_parameter

---

source

SampaddingConv1D_BN

 SampaddingConv1D_BN (in_channels, out_channels, kernel_size)

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

bs = 16
c_in = 3
seq_len = 12
c_out = 2
xb = torch.rand(bs, c_in, seq_len)
m = create_model(OmniScaleCNN, c_in, c_out, seq_len)
test_eq(OmniScaleCNN(c_in, c_out, seq_len)(xb).shape, [bs, c_out])
m

OmniScaleCNN(
  (net): Sequential(
    (0): build_layer_with_layer_parameter(
      (conv_list): ModuleList(
        (0): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 0), value=0)
          (conv1d): Conv1d(3, 56, kernel_size=(1,), stride=(1,))
          (bn): BatchNorm1d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (1): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 1), value=0)
          (conv1d): Conv1d(3, 56, kernel_size=(2,), stride=(1,))
          (bn): BatchNorm1d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (2): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(1, 1), value=0)
          (conv1d): Conv1d(3, 56, kernel_size=(3,), stride=(1,))
          (bn): BatchNorm1d(56, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (1): build_layer_with_layer_parameter(
      (conv_list): ModuleList(
        (0): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 0), value=0)
          (conv1d): Conv1d(168, 227, kernel_size=(1,), stride=(1,))
          (bn): BatchNorm1d(227, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (1): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 1), value=0)
          (conv1d): Conv1d(168, 227, kernel_size=(2,), stride=(1,))
          (bn): BatchNorm1d(227, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (2): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(1, 1), value=0)
          (conv1d): Conv1d(168, 227, kernel_size=(3,), stride=(1,))
          (bn): BatchNorm1d(227, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
    (2): build_layer_with_layer_parameter(
      (conv_list): ModuleList(
        (0): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 0), value=0)
          (conv1d): Conv1d(681, 510, kernel_size=(1,), stride=(1,))
          (bn): BatchNorm1d(510, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (1): SampaddingConv1D_BN(
          (padding): ConstantPad1d(padding=(0, 1), value=0)
          (conv1d): Conv1d(681, 510, kernel_size=(2,), stride=(1,))
          (bn): BatchNorm1d(510, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
    )
  )
  (gap): GAP1d(
    (gap): AdaptiveAvgPool1d(output_size=1)
    (flatten): Flatten(full=False)
  )
  (hidden): Linear(in_features=1020, out_features=2, bias=True)
)