A few fixes

CHANGES.md

@@ -10,6 +10,10 @@
 
 ## Fixes
 
+- Potential bug in convolutional models where checking the out size of the head would affect the batchnorm averaging
+- Potential bug in `plot_sample_pred` to do with bin ranges
+
+
 ## Changes
 
 ## Depreciations

examples/RNNs_CNNs_and_GNNs_for_matrix_data.ipynb

@@ -4290,7 +4290,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.3"
+   "version": "3.6.5"
   },
   "toc": {
    "base_numbering": 1,

lumin/nn/callbacks/abs_callback.py

@@ -5,19 +5,19 @@
 
 class AbsCallback(ABC):
     '''Abstract callback class for typing'''
-    def __init__(self):                        pass
-    def set_model(self, **kargs):              pass
-    def set_plot_settings(self, **kargs):      pass
-    def on_train_begin(self, **kargs):         pass  
-    def on_train_end(self,   **kargs):         pass  
-    def on_epoch_begin(self, **kargs):         pass
-    def on_epoch_end(self,   **kargs):         pass
-    def on_batch_begin(self, **kargs):         pass
-    def on_batch_end(self,   **kargs):         pass
-    def on_eval_begin(self, **kargs):          pass
-    def on_eval_end(self,   **kargs):          pass
-    def on_backwards_begin(self, **kargs):     pass
-    def on_backwards_end(self,   **kargs):     pass
-    def on_pred_begin(self, **kargs):          pass
-    def on_pred_end(self,   **kargs):          pass
+    def __init__(self):                    pass
+    def set_model(self, **kargs):          pass
+    def set_plot_settings(self, **kargs):  pass
+    def on_train_begin(self, **kargs):     pass  
+    def on_train_end(self,   **kargs):     pass  
+    def on_epoch_begin(self, **kargs):     pass
+    def on_epoch_end(self,   **kargs):     pass
+    def on_batch_begin(self, **kargs):     pass
+    def on_batch_end(self,   **kargs):     pass
+    def on_eval_begin(self, **kargs):      pass
+    def on_eval_end(self,   **kargs):      pass
+    def on_backwards_begin(self, **kargs): pass
+    def on_backwards_end(self,   **kargs): pass
+    def on_pred_begin(self, **kargs):      pass
+    def on_pred_end(self,   **kargs):      pass
 

lumin/nn/models/blocks/head.py

@@ -685,7 +685,10 @@ def check_out_sz(self) -> int:
         '''
 
         x = torch.rand((1, self.n_fpv,self.n_v))
+        training = self.training
+        self.eval()
         x = self.forward(x)
+        if training: self.train()
         return x.size(-1)
 
     def get_conv1d_block(self, in_c:int, out_c:int, kernel_sz:int, padding:Union[int,str]='auto', stride:int=1,act:str='relu', bn:bool=False) -> Conv1DBlock:
@@ -941,7 +944,10 @@ def check_out_sz(self) -> int:
         '''
 
         x = torch.rand((1,self.n_v, self.n_fpv))
+        training = self.training
+        self.eval()
         x = self.forward(x)
+        if training: self.train()
         return x.size(-1)
 
 

lumin/nn/models/initialisations.py

@@ -21,14 +21,14 @@ def lookup_normal_init(act:str, fan_in:Optional[int]=None, fan_out:Optional[int]
         Callable to initialise weight tensor
     '''
 
-    if act == 'relu':       return partial(nn.init.kaiming_normal_, nonlinearity='relu')
-    if act == 'prelu':      return partial(nn.init.kaiming_normal_, nonlinearity='relu')
+    if act == 'relu':       return partial(nn.init.kaiming_normal_, nonlinearity='relu', a=0)
+    if act == 'prelu':      return partial(nn.init.kaiming_normal_, nonlinearity='relu', a=0)
     if act == 'selu':       return partial(nn.init.normal_, std=1/np.sqrt(fan_in))
     if act == 'sigmoid':    return nn.init.xavier_normal_
     if act == 'logsoftmax': return nn.init.xavier_normal_
     if act == 'softmax':    return nn.init.xavier_normal_
     if act == 'linear':     return nn.init.xavier_normal_
-    if 'swish' in act:      return partial(nn.init.kaiming_normal_, nonlinearity='relu')
+    if 'swish' in act:      return partial(nn.init.kaiming_normal_, nonlinearity='relu', a=0)
     raise ValueError("Activation not implemented")
 
 
@@ -45,13 +45,13 @@ def lookup_uniform_init(act:str, fan_in:Optional[int]=None, fan_out:Optional[int
         Callable to initialise weight tensor
     '''
 
-    if act == 'relu':       return partial(nn.init.kaiming_uniform_, nonlinearity='relu')
-    if act == 'prelu':      return partial(nn.init.kaiming_uniform_, nonlinearity='relu')
+    if act == 'relu':       return partial(nn.init.kaiming_uniform_, nonlinearity='relu', a=0)
+    if act == 'prelu':      return partial(nn.init.kaiming_uniform_, nonlinearity='relu', a=0)
     if act == 'selu':       return partial(nn.init.uniform_, a=-1/np.sqrt(fan_in), b=1/np.sqrt(fan_in))
     if act == 'sigmoid':    return nn.init.xavier_uniform_
     if act == 'logsoftmax': return nn.init.xavier_uniform_
     if act == 'softmax':    return nn.init.xavier_uniform_
     if act == 'linear':     return nn.init.xavier_uniform_
-    if 'swish' in act:      return partial(nn.init.kaiming_uniform_, nonlinearity='relu')
+    if 'swish' in act:      return partial(nn.init.kaiming_uniform_, nonlinearity='relu', a=0)
     raise ValueError("Activation not implemented")
 

lumin/plotting/results.py

@@ -183,9 +183,9 @@ def plot_sample_pred(df:pd.DataFrame, pred_name:str='pred', targ_name:str='gen_t
         settings: :class:`~lumin.plotting.plot_settings.PlotSettings` class to control figure appearance
     '''
 
-    hist_params = {'range': lim_x, 'bins': bins, 'density': density, 'alpha': 0.8, 'stacked':True, 'rwidth':1.0}
     sig,bkg = (df[targ_name] == 1),(df[targ_name] == 0)
     if not isinstance(bins,list): bins = np.linspace(df[pred_name].min(),df[pred_name].max(), bins if isinstance(bins, int) else 10)
+    hist_params = {'range': lim_x, 'bins': bins, 'density': density, 'alpha': 0.8, 'stacked':True, 'rwidth':1.0}
     sig_samples = _get_samples(df[sig], sample_name, wgt_name)
     bkg_samples = _get_samples(df[bkg], sample_name, wgt_name)
     sample2col = {k: v for v, k in enumerate(bkg_samples)} if settings.sample2col is None else settings.sample2col