diff --git a/docs/requirements.txt b/docs/requirements.txt
index 560a2b3600..d58c576129 100644
--- a/docs/requirements.txt
+++ b/docs/requirements.txt
@@ -1,2 +1,2 @@
-sphinx==2.4.4
+sphinx==3.5.4
 -e git+git://github.com/pytorch/pytorch_sphinx_theme.git#egg=pytorch_sphinx_theme 
diff --git a/docs/source/functional.rst b/docs/source/functional.rst
new file mode 100644
index 0000000000..a40c0941f5
--- /dev/null
+++ b/docs/source/functional.rst
@@ -0,0 +1,25 @@
+.. role:: hidden
+    :class: hidden-section
+
+torchtext.functional
+===========================
+
+.. automodule:: torchtext.functional
+.. currentmodule:: torchtext.functional
+
+to_tensor
+---------
+
+.. autofunction:: to_tensor
+
+
+truncate
+--------
+
+.. autofunction:: truncate
+
+
+add_token
+---------
+
+.. autofunction:: add_token
\ No newline at end of file
diff --git a/docs/source/index.rst b/docs/source/index.rst
index 23b2fb1b52..8a29be9bc3 100644
--- a/docs/source/index.rst
+++ b/docs/source/index.rst
@@ -44,6 +44,9 @@ popular datasets for natural language.
    experimental_vectors
    experimental_vocab
    models_utils
+   transforms
+   functional
+   models
 
 .. automodule:: torchtext
    :members:
diff --git a/docs/source/models.rst b/docs/source/models.rst
new file mode 100644
index 0000000000..500b2a6c7d
--- /dev/null
+++ b/docs/source/models.rst
@@ -0,0 +1,34 @@
+.. role:: hidden
+    :class: hidden-section
+
+torchtext.models
+===========================
+
+.. automodule:: torchtext.models
+.. currentmodule:: torchtext.models
+
+RobertaModelBundle
+------------------
+
+.. autoclass:: RobertaModelBundle
+   :members: transform
+
+   .. automethod:: get_model
+
+XLMR_BASE_ENCODER
+-----------------
+
+.. container:: py attribute
+
+   .. autodata:: XLMR_BASE_ENCODER
+      :no-value:
+
+
+XLMR_LARGE_ENCODER
+------------------
+
+.. container:: py attribute
+
+   .. autodata:: XLMR_LARGE_ENCODER
+      :no-value:
+
diff --git a/docs/source/transforms.rst b/docs/source/transforms.rst
new file mode 100644
index 0000000000..220f18bf34
--- /dev/null
+++ b/docs/source/transforms.rst
@@ -0,0 +1,39 @@
+.. role:: hidden
+    :class: hidden-section
+
+torchtext.transforms
+===========================
+
+.. automodule:: torchtext.transforms
+.. currentmodule:: torchtext.transforms
+
+Transforms are common text transforms. They can be chained together using :class:`torch.nn.Sequential`
+
+SentencePieceTokenizer
+----------------------
+
+.. autoclass:: SentencePieceTokenizer
+
+   .. automethod:: forward
+
+
+VocabTransform
+--------------
+
+.. autoclass:: VocabTransform
+
+   .. automethod:: forward
+
+ToTensor
+--------
+
+.. autoclass:: ToTensor
+
+   .. automethod:: forward
+ 
+LabelToIndex
+------------
+
+.. autoclass:: LabelToIndex
+
+   .. automethod:: forward
diff --git a/torchtext/models/roberta/bundler.py b/torchtext/models/roberta/bundler.py
index bbecc47410..d0ba4b7028 100644
--- a/torchtext/models/roberta/bundler.py
+++ b/torchtext/models/roberta/bundler.py
@@ -23,7 +23,8 @@
 
 @dataclass
 class RobertaModelBundle:
-    """
+    """RobertaModelBundle(_params: torchtext.models.RobertaEncoderParams, _path: Optional[str] = None, _head: Optional[torch.nn.Module] = None, transform: Optional[Callable] = None)
+
     Example - Pretrained encoder
         >>> import torch, torchtext
         >>> xlmr_base = torchtext.models.XLMR_BASE_ENCODER
@@ -66,6 +67,8 @@ class RobertaModelBundle:
     transform: Optional[Callable] = None
 
     def get_model(self, head: Optional[Module] = None, load_weights: bool = True, freeze_encoder: bool = False, *, dl_kwargs=None) -> RobertaModel:
+        r"""get_model(head: Optional[torch.nn.Module] = None, load_weights: bool = True, freeze_encoder: bool = False, *, dl_kwargs=None) -> torctext.models.RobertaModel
+        """
 
         if load_weights:
             assert self._path is not None, "load_weights cannot be True. The pre-trained model weights are not available for the current object"
@@ -108,6 +111,15 @@ def encoderConf(self) -> RobertaEncoderConf:
                       )
 )
 
+XLMR_BASE_ENCODER.__doc__ = (
+    '''
+    XLM-R Encoder with base configuration
+
+    Please refer to :func:`torchtext.models.RobertaModelBundle` for the usage.
+    '''
+)
+
+
 XLMR_LARGE_ENCODER = RobertaModelBundle(
     _path=urljoin(_TEXT_BUCKET, "xlmr.large.encoder.pt"),
     _encoder_conf=RobertaEncoderConf(vocab_size=250002, embedding_dim=1024, ffn_dimension=4096, num_attention_heads=16, num_encoder_layers=24),
@@ -116,3 +128,11 @@ def encoderConf(self) -> RobertaEncoderConf:
                       spm_model_path=urljoin(_TEXT_BUCKET, "xlmr.sentencepiece.bpe.model"),
                       )
 )
+
+XLMR_LARGE_ENCODER.__doc__ = (
+    '''
+    XLM-R Encoder with Large configuration
+
+    Please refer to :func:`torchtext.models.RobertaModelBundle` for the usage.
+    '''
+)
diff --git a/torchtext/transforms.py b/torchtext/transforms.py
index d13f690bc8..cf43e40f4d 100644
--- a/torchtext/transforms.py
+++ b/torchtext/transforms.py
@@ -63,7 +63,7 @@ class VocabTransform(Module):
         >>> jit_vocab_transform = torch.jit.script(vocab_transform)
     """
 
-    def __init__(self, vocab):
+    def __init__(self, vocab: Vocab):
         super().__init__()
         assert isinstance(vocab, Vocab)
         self.vocab = vocab