diff --git a/bdpy/dl/torch/models.py b/bdpy/dl/torch/models.py
index 8e5954c1..5ab0d61a 100644
--- a/bdpy/dl/torch/models.py
+++ b/bdpy/dl/torch/models.py
@@ -1,5 +1,6 @@
 """Model definitions."""
 
+from __future__ import annotations
 
 from typing import Dict, Union, Optional, Sequence
 
@@ -93,7 +94,7 @@ def _parse_layer_name(model: nn.Module, layer_name: str) -> nn.Module:
         Network model.
     layer_name : str
         Layer name. It accepts the following formats: 'layer_name',
-        'layer_name[index]', 'parent_name.child_name', and combinations of them.
+        '[index]', 'parent_name.child_name', and combinations of them.
 
     Returns
     -------
@@ -123,18 +124,19 @@ def _get_value_by_indices(array, indices):
         model = _parse_layer_name(model, top_most_layer_name)
         return _parse_layer_name(model, child_layer_name)
 
-    # parse layer name having index (e.g., 'features[0]', 'backbone[0][1]')
-    pattern = re.compile(r'^(?P<layer_name>\w+)(?P<index>(\[(\d+)\])+)$')
+    # parse layer name having index (e.g., '[0]', 'features[0]', 'backbone[0][1]')
+    pattern = re.compile(r'^(?P<layer_name>[a-zA-Z_]+[a-zA-Z0-9_]*)?(?P<index>(\[(\d+)\])+)$')
     m = pattern.match(layer_name)
     if m is not None:
-        layer_name = m.group('layer_name')
+        layer_name: str | None = m.group('layer_name')  # NOTE: layer_name can be None
         index_str = m.group('index')
 
         indeces = re.findall(r'\[(\d+)\]', index_str)
         indeces = [int(i) for i in indeces]
 
-        if hasattr(model, layer_name):
-            return _get_value_by_indices(getattr(model, layer_name), indeces)
+        if isinstance(layer_name, str) and hasattr(model, layer_name):
+            model = getattr(model, layer_name)
+        return _get_value_by_indices(model, indeces)
 
     raise ValueError(
         f"Invalid layer name: '{layer_name}'. Either the syntax of '{layer_name}' is not supported, "
diff --git a/test/dl/torch/test_models.py b/test/dl/torch/test_models.py
index 6e217d88..dbb33520 100644
--- a/test/dl/torch/test_models.py
+++ b/test/dl/torch/test_models.py
@@ -6,6 +6,13 @@
 from bdpy.dl.torch import models
 
 
+def _removeprefix(text: str, prefix: str) -> str:
+    """Remove prefix from text. (Workaround for Python 3.8)"""
+    if text.startswith(prefix):
+        return text[len(prefix):]
+    return text
+
+
 class MockModule(nn.Module):
     def __init__(self):
         super(MockModule, self).__init__()
@@ -69,6 +76,22 @@ def test_parse_layer_name(self):
         self.assertRaises(
             ValueError, models._parse_layer_name, self.mock, 'layers["key"]')
 
+    def test_parse_layer_name_for_sequential(self):
+        """Test _parse_layer_name for nn.Sequential.
+
+        nn.Sequential is a special case because the submodules are directly
+        accessible like a list. For example, `model[0]` will return the first
+        module in the model.
+        """
+        sequential_module = self.mock.layers
+        accessors = [accessor for accessor in self.accessors if accessor['name'].startswith('layers')]
+        for accessor in accessors:
+            accsessor_key = _removeprefix(accessor['name'], 'layers')
+            layer = models._parse_layer_name(sequential_module, accsessor_key)
+            self.assertIsInstance(layer, accessor['type'])
+            for attr, value in accessor['attrs'].items():
+                self.assertEqual(getattr(layer, attr), value)
+
 
 class TestVGG19(unittest.TestCase):
     def setUp(self):