Hparams passing with user warnings

audio_toolset/audio_io.py

@@ -3,12 +3,13 @@ import torch
 from scipy.signal import butter, lfilter
 
 from ml_lib.modules.utils import AutoPad
-
+import numpy as np
 
 def butter_lowpass(cutoff, sr, order=5):
     nyq = 0.5 * sr
     normal_cutoff = cutoff / nyq
-    b, a = butter(order, normal_cutoff, btype='low', analog=False)
+    # noinspection PyTupleAssignmentBalance
+    b, a = butter(order, normal_cutoff, btype='low', analog=False, output='ba')
     return b, a
 
 
@@ -57,18 +58,19 @@ class NormalizeMelband(object):
         return x
 
 
-class AutoPadTransform(object):
+class AutoPadToShape(object):
-    def __init__(self, **kwargs):
+    def __init__(self, shape):
-        self.__dict__.update(kwargs)
+        self.shape = shape
-        self.padder = AutoPad()
 
-    def __call__(self, y):
+    def __call__(self, x):
-        if not torch.is_tensor(y):
+        if not torch.is_tensor(x):
-            y = torch.as_tensor(y)
+            x = torch.as_tensor(x)
-        return self.padder(y)
+        embedding = torch.zeros(self.shape)
+        embedding[: x.shape] = x
+        return embedding
 
     def __repr__(self):
-        return 'AutoPadTransform()'
+        return f'AutoPadTransform({self.shape})'
 
 
 class Melspectogram(object):

modules/blocks.py

@@ -1,4 +1,5 @@
 from typing import Union
+import warnings
 
 import torch
 from torch import nn
@@ -19,9 +20,9 @@ class ConvModule(nn.Module):
 
     def __init__(self, in_shape, conv_filters, conv_kernel, activation: nn.Module = nn.ELU, pooling_size=None,
                  use_bias=True, use_norm=False, dropout: Union[int, float] = 0,
-                 conv_class=nn.Conv2d, conv_stride=1, conv_padding=0):
+                 conv_class=nn.Conv2d, conv_stride=1, conv_padding=0, **kwargs):
         super(ConvModule, self).__init__()
-
+        warnings.warn(f'The following arguments have been ignored: \n {list(kwargs.keys())}')
         # Module Parameters
         self.in_shape = in_shape
         in_channels, height, width = in_shape[0], in_shape[1], in_shape[2]

modules/utils.py

@@ -11,8 +11,16 @@ import pytorch_lightning as pl
 
 # Utility - Modules
 ###################
+class F_x(object):
+    def __init__(self):
+        pass
+
+    def __call__(self, x):
+        return x
 
 
+# Utility - Modules
+###################
 class Flatten(nn.Module):
 
     @property

utils/config.py

@@ -116,7 +116,7 @@ class Config(ConfigParser, ABC):
     def build_model(self):
         return self.model_class(self.model_paramters)
 
-    def build_and_init_model(self, weight_init_function):
+    def build_and_init_model(self, in_shape, weight_init_function):
         model = self.build_model()
         model.init_weights(weight_init_function)
         return model

utils/logging.py

@@ -37,8 +37,8 @@ class Logger(LightningLoggerBase):
 
     @property
     def outpath(self):
-        # ToDo: Add further path modification such as dataset config etc.
+        # FIXME: Move this out of here, this is not the right place to do this!!!
-        return Path(self.config.train.outpath) / self.config.data.mode
+        return Path(self.config.train.outpath) / self.config.model.type
 
     def __init__(self, config: Config):
         """