Update x.py

x.py

@@ -94,15 +94,15 @@ class DynamicModel(nn.Module):
             layers.append(nn.LeakyReLU(negative_slope=0.01, inplace=True))
         elif activation == 'elu':
             layers.append(nn.ELU(alpha=1.0, inplace=True))
-        if dropout := lp.get('dropout', 0.0):
+        if dropout := lp.get('dropout', 0.1):
             layers.append(nn.Dropout(p=dropout))
-        if lp.get('memory_augmentation', False):
+        if lp.get('memory_augmentation', True):
             layers.append(MemoryAugmentationLayer(lp['output_size']))
-        if lp.get('hybrid_attention', False):
+        if lp.get('hybrid_attention', True):
             layers.append(HybridAttentionLayer(lp['output_size']))
-        if lp.get('dynamic_flash_attention', False):
+        if lp.get('dynamic_flash_attention', True):
             layers.append(DynamicFlashAttentionLayer(lp['output_size']))
-        if lp.get('magic_state', False):
+        if lp.get('magic_state', True):
             layers.append(MagicStateLayer(lp['output_size']))
         return nn.Sequential(*layers)
 
@@ -118,7 +118,7 @@ class DynamicModel(nn.Module):
 
 def parse_xml_file(file_path):
     tree, root, layers = ET.parse(file_path), ET.parse(file_path).getroot(), []
-    for layer in root.findall('.//layer'):
+    for layer in root.findall('.//label'):
         lp = {
             'input_size': int(layer.get('input_size', 128)),
             'output_size': int(layer.get('output_size', 256)),