docs: add the docstrings for v1.0.0 (#656)

* add v2 flux examples

* add the docs

* add docs

* update

* finished ops

* add ops

* update

* update

* update

* update

* update

* update

* update

* update docstrings

* update

* update

* update

* update

* update

* update

* update

* finished the api docs

* update

* update

docs/source/developer/docstring.rst

+Writing Docstrings
+==================
+
+Use this guide to write clear, consistent Python docstrings for Nunchaku.
+Follow the `NumPy style guide <https://numpydoc.readthedocs.io/en/latest/format.html>`__, and always specify variable shapes, dtypes, and notation.
+The docstring should be concise and informative.
+
+Docstring Format
+----------------
+
+A standard docstring should look like:
+
+.. code-block:: python
+
+    """
+    Short summary of what the function or class does.
+
+    (Optional) Extended description.
+
+    Parameters
+    ----------
+    param1 : type
+        Description.
+    param2 : type, optional
+        Description. Default is ...
+    param3 : array-like, shape (n, m), dtype float
+        Example of shape and dtype notation.
+
+    Returns
+    -------
+    out1 : type
+        Description.
+    out2 : type
+        Description.
+
+    Raises
+    ------
+    ValueError
+        When this exception is raised.
+
+    See Also
+    --------
+    other_function : brief description
+
+    Notes
+    -----
+    Additional details or references.
+
+    Examples
+    --------
+    >>> result = func(1, 2)
+    >>> print(result)
+    3
+    """
+
+Guidelines
+----------
+
+- Use triple double quotes (``"""``) for all docstrings.
+- Every public module, class, method, and function must have a docstring.
+- The first line is a concise summary.
+- Use sections in this order (as needed): ``Parameters``, ``Returns``, ``Raises``, ``See Also``, ``Notes``, ``Examples``.
+
+Shapes, Dtypes, and Notation
+----------------------------
+
+- Always specify expected shape and dtype for tensors/arrays.
+- Use plain text for shapes (not LaTeX/math symbols).
+- Use clear, single-letter or descriptive names for shape dimensions (e.g., `B` for batch size).
+- Define all shape symbols in a `Notes` section.
+
+**Example:**
+
+.. code-block:: python
+
+    def forward(
+        self,
+        x: torch.Tensor,
+        mask: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        """
+        Forward pass of the model.
+
+        Parameters
+        ----------
+        x : torch.Tensor, shape (B, C, H, W), dtype float32
+            Input image tensor.
+        mask : Optional[torch.Tensor], shape (B, 1, H, W), dtype bool
+            Optional mask.
+
+        Returns
+        -------
+        out : torch.Tensor, shape (B, num_classes), dtype float32
+            Output logits.
+
+        Raises
+        ------
+        ValueError
+            If input shapes are incompatible.
+
+        Notes
+        -----
+        Notations:
+        - B: batch size
+        - C: channels
+        - H: height
+        - W: width
+        - num_classes: number of output classes
+
+        Examples
+        --------
+        >>> x = torch.randn(8, 3, 224, 224)
+        >>> out = model.forward(x)
+        """
+        ...
+
+Best Practices
+--------------
+
+Writing Tips
+~~~~~~~~~~~~
+
+- **Be concise and clear.** Start with a short summary describing what the function or class does.
+- **Document all parameters and return values.** Indicate if a parameter can be `None`.
+- **Include an** ``Examples`` **section** to demonstrate typical usage.
+- **List all possible exceptions in a** ``Raises`` **section.**
+- **Use a** ``Notes`` **section** to define shape symbols and explain special behaviors.
+- **Add a** ``See Also`` **section** for related functions or methods.
+
+Useful Prompts
+~~~~~~~~~~~~~~
+
+.. code-block:: text
+
+   Improve the writing of the docstring according to this guide. Be concise. Organize my comments clearly.
+
+   Write the docstring for this module (every class and functions) according to this guide. Each function and class that can be shwon properly and beautifully in sphinx html. Be concise. Organize my comments clearly.
+
+For further questions or formatting help, refer to existing Nunchaku code or ask in the developer chat.

docs/source/index.rst

@@ -16,6 +16,7 @@ Check out `DeepCompressor <github_deepcompressor_>`_ for the quantization librar
     :caption: Usage Tutorials
 
     usage/basic_usage.rst
+    usage/qwenimage.rst
     usage/lora.rst
     usage/kontext.rst
     usage/controlnet.rst
@@ -48,3 +49,4 @@ Check out `DeepCompressor <github_deepcompressor_>`_ for the quantization librar
 
     faq/faq.rst
     developer/contribution_guide.rst
+    developer/docstring.rst

docs/source/links/huggingface.txt

@@ -9,3 +9,5 @@
 .. _hf_depth_anything: https://huggingface.co/LiheYoung/depth-anything-large-hf
 .. _hf_nunchaku_wheels: https://huggingface.co/nunchaku-tech/nunchaku
 .. _hf_ip-adapterv2: https://huggingface.co/XLabs-AI/flux-ip-adapter-v2
+.. _hf_qwen-image: https://huggingface.co/Qwen/Qwen-Image
+.. _hf_qwen-image-lightning: https://huggingface.co/lightx2v/Qwen-Image-Lightning

docs/source/links/paper.txt

 .. _paper_svdquant: http://arxiv.org/abs/2411.05007
 .. _paper_pulid: https://arxiv.org/abs/2404.16022
+.. _paper_awq: https://arxiv.org/abs/2306.00978

docs/source/python_api/nunchaku.models.attention.rst

+nunchaku.models.attention
+=============================
+
+.. automodule:: nunchaku.models.attention
+   :members:
+   :undoc-members:
+   :show-inheritance:
+   :private-members:

docs/source/python_api/nunchaku.models.attention_processors.flux.rst

+nunchaku.models.attention_processors.flux
+=========================================
+
+.. automodule:: nunchaku.models.attention_processors.flux
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.models.attention_processors.qwenimage.rst

+nunchaku.models.attention_processors.qwenimage
+==============================================
+
+.. automodule:: nunchaku.models.attention_processors.qwenimage
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.models.attention_processors.rst

+nunchaku.models.attention_processors
+====================================
+
+.. toctree::
+   :maxdepth: 4
+
+   nunchaku.models.attention_processors.flux
+   nunchaku.models.attention_processors.qwenimage

docs/source/python_api/nunchaku.models.embeddings.rst

+nunchaku.models.embeddings
+==========================
+
+.. automodule:: nunchaku.models.embeddings
+   :members:
+   :undoc-members:
+   :show-inheritance:
+   :private-members:

docs/source/python_api/nunchaku.models.linear.rst

+nunchaku.models.linear
+======================
+
+.. automodule:: nunchaku.models.linear
+   :members:
+   :undoc-members:
+   :show-inheritance:
+   :private-members:

docs/source/python_api/nunchaku.models.normalization.rst

+nunchaku.models.normalization
+=============================
+
+.. automodule:: nunchaku.models.normalization
+   :members:
+   :undoc-members:
+   :show-inheritance:
+   :private-members:

docs/source/python_api/nunchaku.models.rst

@@ -6,6 +6,12 @@ nunchaku.models
 
    nunchaku.models.transformers
    nunchaku.models.text_encoders
+   nunchaku.models.linear
+   nunchaku.models.attention
+   nunchaku.models.attention_processors
+   nunchaku.models.normalization
+   nunchaku.models.embeddings
    nunchaku.models.pulid
    nunchaku.models.ip_adapter
    nunchaku.models.safety_checker
+   nunchaku.models.utils

docs/source/python_api/nunchaku.models.transformers.rst

@@ -5,5 +5,7 @@ nunchaku.models.transformers
    :maxdepth: 4
 
    nunchaku.models.transformers.transformer_flux
+   nunchaku.models.transformers.transformer_flux_v2
+   nunchaku.models.transformers.transformer_qwenimage
    nunchaku.models.transformers.transformer_sana
    nunchaku.models.transformers.utils

docs/source/python_api/nunchaku.models.transformers.transformer_flux_v2.rst

+nunchaku.models.transformers.transformer\_flux\_v2
+==================================================
+
+.. automodule:: nunchaku.models.transformers.transformer_flux_v2
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.models.transformers.transformer_qwenimage.rst

+nunchaku.models.transformers.transformer\_qwenimage
+===================================================
+
+.. automodule:: nunchaku.models.transformers.transformer_qwenimage
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.models.utils.rst

+nunchaku.models.utils
+=====================
+
+.. automodule:: nunchaku.models.utils
+   :members:
+   :undoc-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.ops.fused.rst

+nunchaku.ops.fused
+==================
+
+.. automodule:: nunchaku.ops.fused
+   :members:
+   :undoc-members:
+   :private-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.ops.gemm.rst

+nunchaku.ops.gemm
+=================
+
+.. automodule:: nunchaku.ops.gemm
+   :members:
+   :undoc-members:
+   :private-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.ops.gemv.rst

+nunchaku.ops.gemv
+=================
+
+.. automodule:: nunchaku.ops.gemv
+   :members:
+   :undoc-members:
+   :private-members:
+   :show-inheritance:

docs/source/python_api/nunchaku.ops.quantize.rst

+nunchaku.ops.quantize
+=====================
+
+.. automodule:: nunchaku.ops.quantize
+   :members:
+   :undoc-members:
+   :private-members:
+   :show-inheritance: