Update docs

docs/Project.toml

@@ -3,8 +3,6 @@ Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
 DocumenterMarkdown = "997ab1e6-3595-5248-9280-8efb232c3433"
 Flux = "587475ba-b771-5e3f-ad9e-33799f191a9c"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
-LuxCore = "bb33d45b-7691-41d6-9220-0943567d0623"
 Lux = "b2108857-7c20-44ae-9111-449ecde12c47"
-LuxLib = "82251201-b29d-42c6-8e01-566dec8acb11"
 Optimisers = "3bd65402-5787-11e9-1adc-39752487f4e2"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"

docs/make.jl

@@ -1,22 +1,14 @@
-using Documenter, DocumenterMarkdown, LuxCore, Lux, LuxLib, Pkg
+using Documenter, DocumenterMarkdown, Lux, Pkg
 
 import Flux  # Load weak dependencies
 
 deployconfig = Documenter.auto_detect_deploy_system()
 Documenter.post_status(deployconfig; type="pending", repo="github.com/avik-pal/Lux.jl.git")
 
 makedocs(; sitename="Lux", authors="Avik Pal et al.", clean=true, doctest=true,
-         modules=[Lux, LuxLib, LuxCore],
-         strict=[
-             :doctest,
-             :linkcheck,
-             :parse_error,
-             :example_block,
-             # Other available options are
-             # :autodocs_block, :cross_references, :docs_block, :eval_block, :example_block,
-             # :footnote, :meta_block, :missing_docs, :setup_block
-         ], checkdocs=:all, format=Markdown(), draft=false,
-         build=joinpath(@__DIR__, "docs"))
+         modules=[Lux],
+         strict=[:doctest, :linkcheck, :parse_error, :example_block, :missing_docs],
+         checkdocs=:all, format=Markdown(), draft=false, build=joinpath(@__DIR__, "docs"))
 
 deploydocs(; repo="github.com/avik-pal/Lux.jl.git", push_preview=true,
            deps=Deps.pip("mkdocs", "pygments", "python-markdown-math", "mkdocs-material",

docs/mkdocs.yml

@@ -113,16 +113,6 @@ nav:
       - "Utilities": "api/utilities.md"
       - "Flux & Lux InterOp": "api/flux2lux.md"
       - "Experimental": "api/contrib.md"
-  - "Frameworks":
-      - "Boltz": "lib/Boltz/index.md"
-      - "FluxMPI": "https://avik-pal.github.io/FluxMPI.jl/dev/"
-      - "NNlib": "https://fluxml.ai/Flux.jl/stable/models/nnlib/"
-      - "LuxLib":
-        - "Introduction": "lib/LuxLib/index.md"
-        - "API Reference": "lib/LuxLib/api.md"
-      - "LuxCore":
-        - "Introduction": "lib/LuxCore/index.md"
-        - "API Reference": "lib/LuxCore/api.md"
   - "Development Documentation":
       - "Style Guide": "devdocs/style_guide.md"
       - "Layer Implementation": "devdocs/layer_implementation.md"

docs/src/api/contrib.md

@@ -1,3 +1,5 @@
+# Experimental Features
+
 ```@meta
 CurrentModule = Lux
 ```
@@ -11,6 +13,12 @@ All features listed on this page are **experimental** which means:
      experimental sooner.
   4. None of the features are exported.
 
+## Index
+
+```@index
+Pages = ["contrib.md"]
+```
+
 ## Training
 
 !!! note
@@ -70,9 +78,3 @@ Lux.@layer_map
 ```@docs
 Lux.share_parameters
 ```
-
-## Index
-
-```@index
-Pages = ["contrib.md"]
-```

docs/src/api/core.md

@@ -1,8 +1,10 @@
+# LuxCore.jl
+
 ```@meta
 CurrentModule = Lux
 ```
 
-The documentation for this page has been moved to [LuxCore.jl](../lib/LuxCore/api.md).
+The documentation for this page has been moved to [LuxCore.jl](https://luxdl.github.io/LuxCore.jl/dev).
 However, this the functionality has in **no way been deprecated** and there is no plan
 to deprecated access to these functionalities in the future. The change was made merely to
 allow lighter dependencies for users extending `Lux.jl`.

docs/src/api/flux2lux.md

@@ -1,10 +1,18 @@
+# Flux Models to Lux Models
+
 ```@meta
 CurrentModule = Lux
 ```
 
 Accessing these functions require manually loading `Flux`, i.e., `using Flux` must be
 present somewhere in the code for these to be used.
 
+## Index
+
+```@index
+Pages = ["flux2lux.md"]
+```
+
 ## Functions
 
 ```@docs
@@ -16,9 +24,3 @@ transform
 ```@docs
 FluxLayer
 ```
-
-## Index
-
-```@index
-Pages = ["flux2lux.md"]
-```

docs/src/api/functional.md

@@ -1,12 +1,16 @@
+# Functional Layers
+
 ```@meta
 CurrentModule = Lux
 ```
 
-## Functional Layers
+!!! tip
+    It is recommended to use [LuxLib.jl](https://luxdl.github.io/LuxLib.jl/dev) for these
+    functionality.
 
 !!! note
-    These functions expose the backend of `Lux.jl`. In the long-term we plan to move these
-    into NNlib
+    These functions expose the backend of `Lux.jl`. These will be removed in future
+    releases
 
 ```@docs
 Lux.dropout

docs/src/api/layers.md

@@ -1,7 +1,15 @@
+# Layers
+
 ```@meta
 CurrentModule = Lux
 ```
 
+## Index
+
+```@index
+Pages = ["layers.md"]
+```
+
 ## Containers
 
 ```@docs
@@ -86,9 +94,3 @@ WeightNorm
 PixelShuffle
 Upsample
 ```
-
-## Index
-
-```@index
-Pages = ["layers.md"]
-```

docs/src/api/utilities.md

@@ -1,7 +1,15 @@
+# Utilities
+
 ```@meta
 CurrentModule = Lux
 ```
 
+## Index
+
+```@index
+Pages = ["utilities.md"]
+```
+
 ## Data Transfer
 
 ```@docs
@@ -35,12 +43,6 @@ Lux.replicate
 
 ## Truncated Stacktraces
 
-```docs
+```@docs
 Lux.disable_stacktrace_truncation!
 ```
-
-## Index
-
-```@index
-Pages = ["utilities.md"]
-```

docs/src/devdocs/style_guide.md

@@ -19,12 +19,12 @@ We do have automatic formatter, which opens PR after fixing common style issues,
 
 ## Code Styling
 
-* Keyword Arguments must be separated using a semicolon `;`
-* Functions must use `return`. Returning the last value is quite ambiguous -- did the author
+- Keyword Arguments must be separated using a semicolon `;`
+- Functions must use `return`. Returning the last value is quite ambiguous -- did the author
   actually want it returned?
-* Format docstrings as you would format regular code. If the docstring constains LaTeX in
+- Format docstrings as you would format regular code. If the docstring constains LaTeX in
   multiple lines, use `math` block.
-* No avoiding multiply symbol -- so `2x` is invalid instead do it like other languages
+- No avoiding multiply symbol -- so `2x` is invalid instead do it like other languages
   `2 * x`.
 
 ## Testing
@@ -33,17 +33,17 @@ We do have automatic formatter, which opens PR after fixing common style issues,
     Unfortunately we haven't yet tested all the functionality in the base library using
     these guidelines.
 
-* The file structure of the `test` folder should mirror that of the `src` folder. Every file
+- The file structure of the `test` folder should mirror that of the `src` folder. Every file
   in src should have a complementary file in the test folder, containing tests relevant to
   that file's contents.
 
-* Add generic utilities for testing in `test/test_utils.jl` and include them in the relevant
+- Add generic utilities for testing in `test/test_utils.jl` and include them in the relevant
   files.
 
-* Use [JET.jl](https://aviatesk.github.io/JET.jl/dev/) to test for dynamic dispatch in the
+- Use [JET.jl](https://aviatesk.github.io/JET.jl/dev/) to test for dynamic dispatch in the
   functionality you added, specifically use `run_JET_tests` from `test/test_utils.jl`.
 
-* **Always** test for gradient correctness. Zygote can be notorious for incorrect gradients,
+- **Always** test for gradient correctness. Zygote can be notorious for incorrect gradients,
   so add tests using `test_gradient_correctness_fdm` for finite differencing or use any
   other AD framework and tally the results.
 
@@ -57,7 +57,6 @@ using `NNlib.jl` get to benefit from these fixes.
 Similarly, if a bug comes to the forefront from one of the backend packages, make sure to 
 open a corresponding issue there to ensure they are appropriately tracked.
 
-
 ## Mutability
 
 This is strictly enforced, i.e. all layers/functions provided as part of the external API
@@ -71,13 +70,13 @@ away those optimizations (these can be tested via `Zygote.@code_ir`).
 
 ### Writing efficient non-branching code to make Zygote happy
 
-* Rely on `@generated` functions to remove **most** runtime branching. Certain examples:
-  * Layers behaving differently during training and inference -- we know at compile-time
+- Rely on `@generated` functions to remove **most** runtime branching. Certain examples:
+  - Layers behaving differently during training and inference -- we know at compile-time
     whether a layer is being run in training/inference mode via `istraining(st)`.
-  * Composite Layers relying on a variable number of internal layers -- Again we know the
+  - Composite Layers relying on a variable number of internal layers -- Again we know the
     length of the number of internal layers at compile time. Hence we can manually unroll
     the loops. See [`Parallel`](@ref), [`Chain`](@ref), etc.
-* Pass around `Val` in state. `Flux.jl` sets `training` to be `(:auto, true, false)`. Hence,
+- Pass around `Val` in state. `Flux.jl` sets `training` to be `(:auto, true, false)`. Hence,
   which branch will be evaluated, will have to be determined at runtime time (*bad*).
   Instead if we pass `Val(true)`, we will be able to specialize functions directly based on
   `true`, `false`, etc. ensuring there is no runtime cost for these operations.
@@ -93,14 +92,14 @@ modify their code for upcoming releases.
 process of deprecating functionalities in Julia packages. We follow the same process. Some
 additional guidelines are:
 
-  - Add tests using `Test.@test_deprecated` to ensure that deprecations are indeed working
-    as expected.
-  - Add a warning to the documentation about deprecations (and how to use the new
-    recommended functionality).
-  - Add `# Deprecated Functionality (Remove in <VERSION NUMBER>)` before the tests and
-    deprecated functionality not placed in `src/deprecated.jl` (like kwarg deprecations).
-    This makes it easier to search and delete the functionalities before making a breaking
-    release.
+- Add tests using `Test.@test_deprecated` to ensure that deprecations are indeed working
+  as expected.
+- Add a warning to the documentation about deprecations (and how to use the new
+  recommended functionality).
+- Add `# Deprecated Functionality (Remove in <VERSION NUMBER>)` before the tests and
+  deprecated functionality not placed in `src/deprecated.jl` (like kwarg deprecations).
+  This makes it easier to search and delete the functionalities before making a breaking
+  release.
 
 ## Documentation
 
@@ -127,15 +126,15 @@ equations being used, i.e. equations and the internal code must be consistent.
 Next, we will have certain subsections (though all of them might not be necessary for all
 layers).
 
-* **Arguments**: This section should be present unless the layer is constructed without any
+- **Arguments**: This section should be present unless the layer is constructed without any
   arguments (See [`NoOpLayer`](@ref)). All the arguments and their explicit constraints must
   be explained.
-  * It is recommended to separate out the Keyword Arguments in their own section.
-* **Inputs**: This section should always be present. List out the requirements `x` needs to
+  - It is recommended to separate out the Keyword Arguments in their own section.
+- **Inputs**: This section should always be present. List out the requirements `x` needs to
   satisfy. (Don't write about `ps` and `st` since that is expected by default.)
-* **Returns**: What will the layer return? We know the second element will be a state but is
+- **Returns**: What will the layer return? We know the second element will be a state but is
   that updated in any form or not? 
-* **Parameters**: What are the properties of the NamedTuple returned from
+- **Parameters**: What are the properties of the NamedTuple returned from
   `initialparameters`? Omit if the layer is parameterless.
-* **States**: What are the properties of the NamedTuple returned from `initialstates`? Omit
+- **States**: What are the properties of the NamedTuple returned from `initialstates`? Omit
   if the layer is stateless.