Docs tweaks

Project.toml

@@ -1,7 +1,7 @@
 name = "RecurrentLayers"
 uuid = "78449bcf-6750-4b78-9e82-63d4a1ccdf8c"
 authors = ["Francesco Martinuzzi"]
-version = "0.1.2"
+version = "0.1.3"
 
 [deps]
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"

docs/pages.jl

@@ -4,4 +4,5 @@ pages=[
             "Cells" => "api/cells.md",
             "Cell Wrappers" => "api/wrappers.md",
         ],
+        "Roadmap" => "roadmap.md"
     ]

docs/src/roadmap.md

@@ -0,0 +1,22 @@
+# Roadmap
+This page documents some planned work for RecurrentLayers.jl.
+Future work for this library includes additional cells such as:
+
+ - FastRNNs and FastGRUs (current focus) [arxiv](https://arxiv.org/abs/1901.02358)
+ - Unitary recurrent neural networks [arxiv](https://arxiv.org/abs/1611.00035)
+ - Modern recurrent neural networks such as [LRU](https://arxiv.org/abs/2303.06349) 
+   and [minLSTM/minGRU](https://arxiv.org/abs/2410.01201)
+ - Quasi recurrent neural networks [arxiv](https://arxiv.org/abs/1611.01576)
+
+Additionally, some cell-independent architectures are also planned,
+that expand the ability of recurrent architectures and could theoretically take
+any cell:
+
+ - Clockwork rnns [arxiv](https://arxiv.org/abs/1402.3511)
+ - Phased rnns [arxiv](https://arxiv.org/abs/1610.09513)
+ - Segment rnn [arxiv](https://arxiv.org/abs/2308.11200)
+ - Fast-Slow rnns [arxiv](https://arxiv.org/abs/1705.08639)
+
+An implementation of these ideally would be,
+for example `FastSlow(RNNCell, input_size => hidden_size)`.
+More details on this soon!