Update docs for new layers.

Author: sharvil

docs/pytorch/haste_pytorch.md

@@ -15,7 +15,11 @@ Haste: a fast, simple, and open RNN library.
 
 [`class GRU`](./haste_pytorch/GRU.md): Gated Recurrent Unit layer.
 
+[`class IndRNN`](./haste_pytorch/IndRNN.md): Independently Recurrent Neural Network layer.
+
 [`class LSTM`](./haste_pytorch/LSTM.md): Long Short-Term Memory layer.
 
+[`class LayerNormGRU`](./haste_pytorch/LayerNormGRU.md): Layer Normalized Gated Recurrent Unit layer.
+
 [`class LayerNormLSTM`](./haste_pytorch/LayerNormLSTM.md): Layer Normalized Long Short-Term Memory layer.
 

docs/pytorch/haste_pytorch/GRU.md

@@ -152,7 +152,7 @@ apply(fn)
 
 Applies ``fn`` recursively to every submodule (as returned by ``.children()``)
 as well as self. Typical use includes initializing the parameters of a model
-(see also :ref:`torch-nn-init`).
+(see also :ref:`nn-init-doc`).
 
 #### Args:
 
@@ -341,6 +341,7 @@ Casts all floating point parameters and buffers to float datatype.
 ``` python
 forward(
     input,
+    state=None,
     lengths=None
 )
 ```