More API updates

LuxDL · Jun 11, 2024 · bb41d0c · bb41d0c
1 parent f7f4551
commit bb41d0c
Show file tree

Hide file tree

Showing 5 changed files with 104 additions and 69 deletions.
diff --git a/docs/src/.vitepress/config.mts b/docs/src/.vitepress/config.mts
@@ -206,7 +206,9 @@ export default defineConfig({
                     },
                     {
                         text: 'Domain Specific Modeling', collapsed: false, items: [
-                            { text: 'Boltz', link: '/api/Domain_Specific_Modeling/Boltz' }]
+                            { text: 'Boltz', link: '/api/Domain_Specific_Modeling/Boltz' },
+                            { text: 'Computer Vision', link: '/api/Domain_Specific_Modeling/Boltz_Vision' },
+                            { text: 'Boltz.Layers', link: '/api/Domain_Specific_Modeling/Boltz_Layers' }]
                     },
                     {
                         text: 'Testing Functionality', collapsed: false, items: [

diff --git a/docs/src/api/Domain_Specific_Modeling/Boltz.md b/docs/src/api/Domain_Specific_Modeling/Boltz.md
@@ -9,72 +9,5 @@ Accelerate ⚡ your ML research using pre-built Deep Learning Models with Lux.
 ## Index
 
 ```@index
-Pages = ["Boltz.md"]
-```
-
-## `Layers` API
-
-```@docs
-Layers.ConvBatchNormActivation
-Layers.ClassTokens
-Layers.MultiHeadSelfAttention
-Layers.ViPosEmbedding
-Layers.VisionTransformerEncoder
-```
-
-## Computer Vision Models (`Vision` API)
-
-### Native Lux Models
-
-```@docs
-Vision.VGG
-Vision.VisionTransformer
-```
-
-### Imported from Metalhead.jl
-
-!!! tip
-
-    You need to load `Flux` and `Metalhead` before using these models.
-
-```@docs
-Vision.AlexNet
-Vision.ConvMixer
-Vision.DenseNet
-Vision.GoogLeNet
-Vision.MobileNet
-Vision.ResNet
-Vision.ResNeXt
-```
-
-### Pretrained Models
-
-!!! tip
-
-    Pass `pretrained=true` to the model constructor to load the pretrained weights.
-
-
-| MODEL                     | TOP 1 ACCURACY (%) | TOP 5 ACCURACY (%) |
-| :------------------------ | :----------------: | :----------------: |
-| `AlexNet()`               |       54.48        |       77.72        |
-| `VGG(11)`                 |       67.35        |       87.91        |
-| `VGG(13)`                 |       68.40        |       88.48        |
-| `VGG(16)`                 |       70.24        |       89.80        |
-| `VGG(19)`                 |       71.09        |       90.27        |
-| `VGG(11; batchnorm=true)` |       69.09        |       88.94        |
-| `VGG(13; batchnorm=true)` |       69.66        |       89.49        |
-| `VGG(16; batchnorm=true)` |       72.11        |       91.02        |
-| `VGG(19; batchnorm=true)` |       72.95        |       91.32        |
-
-#### Preprocessing
-
-All the pretrained models require that the images be normalized with the parameters
-`mean = [0.485f0, 0.456f0, 0.406f0]` and `std = [0.229f0, 0.224f0, 0.225f0]`.
-
-## Non-Public API
-
-```@docs
-Boltz._seconddimmean
-Boltz._fast_chunk
-Boltz._flatten_spatial
+Pages = ["Boltz.md", "Boltz_Layers.md", "Boltz_Vision.md"]
 ```
diff --git a/docs/src/api/Domain_Specific_Modeling/Boltz_Layers.md b/docs/src/api/Domain_Specific_Modeling/Boltz_Layers.md
@@ -0,0 +1,36 @@
+```@meta
+CurrentModule = Boltz
+```
+
+# `Boltz.Layers` and `Boltz.Basis` API Reference
+
+## `Layers` API
+
+```@docs
+Layers.ConvBatchNormActivation
+Layers.ConvNormActivation
+Layers.ClassTokens
+Layers.HamiltonianNN
+Layers.MultiHeadSelfAttention
+Layers.MLP
+Layers.TensorProductLayer
+Layers.ViPosEmbedding
+Layers.VisionTransformerEncoder
+```
+
+## Basis Functions
+
+!!! warning
+
+    The function calls for these basis functions should be considered experimental and are
+    subject to change without deprecation. However, the functions themselves are stable
+    and can be freely used in combination with the other Layers and Models.
+
+```@docs
+Basis.Cos
+Basis.Chebychev
+Basis.Fourier
+Basis.Legendre
+Basis.Polynomial
+Basis.Sin
+```
diff --git a/docs/src/api/Domain_Specific_Modeling/Boltz_Private.md b/docs/src/api/Domain_Specific_Modeling/Boltz_Private.md
@@ -0,0 +1,12 @@
+```@meta
+CurrentModule = Boltz
+```
+
+# Boltz.jl Private API
+
+```@docs
+Boltz._seconddimmean
+Boltz._should_type_assert
+Boltz._fast_chunk
+Boltz._flatten_spatial
+```
diff --git a/docs/src/api/Domain_Specific_Modeling/Boltz_Vision.md b/docs/src/api/Domain_Specific_Modeling/Boltz_Vision.md
@@ -0,0 +1,52 @@
+```@meta
+CurrentModule = Boltz
+```
+
+# Computer Vision Models (`Vision` API)
+
+## Native Lux Models
+
+```@docs
+Vision.VGG
+Vision.VisionTransformer
+```
+
+## Imported from Metalhead.jl
+
+!!! tip
+
+    You need to load `Flux` and `Metalhead` before using these models.
+
+```@docs
+Vision.AlexNet
+Vision.ConvMixer
+Vision.DenseNet
+Vision.GoogLeNet
+Vision.MobileNet
+Vision.ResNet
+Vision.ResNeXt
+```
+
+## Pretrained Models
+
+!!! tip
+
+    Pass `pretrained=true` to the model constructor to load the pretrained weights.
+
+
+| MODEL                     | TOP 1 ACCURACY (%) | TOP 5 ACCURACY (%) |
+| :------------------------ | :----------------: | :----------------: |
+| `AlexNet()`               |       54.48        |       77.72        |
+| `VGG(11)`                 |       67.35        |       87.91        |
+| `VGG(13)`                 |       68.40        |       88.48        |
+| `VGG(16)`                 |       70.24        |       89.80        |
+| `VGG(19)`                 |       71.09        |       90.27        |
+| `VGG(11; batchnorm=true)` |       69.09        |       88.94        |
+| `VGG(13; batchnorm=true)` |       69.66        |       89.49        |
+| `VGG(16; batchnorm=true)` |       72.11        |       91.02        |
+| `VGG(19; batchnorm=true)` |       72.95        |       91.32        |
+
+### Preprocessing
+
+All the pretrained models require that the images be normalized with the parameters
+`mean = [0.485f0, 0.456f0, 0.406f0]` and `std = [0.229f0, 0.224f0, 0.225f0]`.