Rename GELU to Gelu (tracel-ai#1311)

* Rename GELU to Gelu This is to follow Rust naming convention
NoahSchiro · Feb 15, 2024 · dfb739c · dfb739c
1 parent d1273d4
commit dfb739c
Show file tree

Hide file tree

Showing 9 changed files with 26 additions and 26 deletions.
diff --git a/README.md b/README.md
@@ -508,7 +508,7 @@ pub struct PositionWiseFeedForward<B: Backend> {
     linear_inner: nn::Linear<B>,
     linear_outer: nn::Linear<B>,
     dropout: nn::Dropout,
-    gelu: nn::GELU,
+    gelu: nn::Gelu,
 }
 
 impl<B: Backend> PositionWiseFeedForward<B> {

diff --git a/burn-book/src/building-blocks/module.md b/burn-book/src/building-blocks/module.md
@@ -14,7 +14,7 @@ pub struct PositionWiseFeedForward<B: Backend> {
     linear_inner: Linear<B>,
     linear_outer: Linear<B>,
     dropout: Dropout,
-    gelu: GELU,
+    gelu: Gelu,
 }
 
 impl<B: Backend> PositionWiseFeedForward<B> {
@@ -113,7 +113,7 @@ Burn comes with built-in modules that you can use to build your own modules.
 | `LayerNorm` | `nn.LayerNorm`                          |
 | `GroupNorm` | `nn.GroupNorm`                          |
 | `Dropout`   | `nn.Dropout`                            |
-| `GELU`      | `nn.GELU`                               |
+| `Gelu`      | `nn.Gelu`                               |
 | `Linear`    | `nn.Linear`                             |
 | `Embedding` | `nn.Embedding`                          |
 | `Relu`      | `nn.ReLU`                               |

diff --git a/burn-core/src/nn/attention/mha.rs b/burn-core/src/nn/attention/mha.rs
@@ -55,7 +55,7 @@ pub struct MultiHeadAttention<B: Backend> {
     value: nn::Linear<B>,
     output: nn::Linear<B>,
     dropout: nn::Dropout,
-    activation: nn::GELU,
+    activation: nn::Gelu,
     n_heads: usize,
     d_k: usize,
     min_float: f64,
@@ -87,7 +87,7 @@ impl MultiHeadAttentionConfig {
             value: linear(self),
             output: linear(self),
             dropout: nn::DropoutConfig::new(self.dropout).init(),
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
             n_heads: self.n_heads,
             d_k: self.d_model / self.n_heads,
             min_float: self.min_float,
@@ -111,7 +111,7 @@ impl MultiHeadAttentionConfig {
             value: linear(self, record.value),
             output: linear(self, record.output),
             dropout: nn::DropoutConfig::new(self.dropout).init(),
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
             n_heads: self.n_heads,
             d_k: self.d_model / self.n_heads,
             min_float: self.min_float,

diff --git a/burn-core/src/nn/gelu.rs b/burn-core/src/nn/gelu.rs
@@ -6,9 +6,9 @@ use crate::tensor::Tensor;
 
 /// Applies the Gaussian Error Linear Units function element-wise.
 #[derive(Module, Clone, Debug, Default)]
-pub struct GELU {}
+pub struct Gelu {}
 
-impl GELU {
+impl Gelu {
     /// Create the module.
     pub fn new() -> Self {
         Self {}

diff --git a/burn-core/src/nn/transformer/pwff.rs b/burn-core/src/nn/transformer/pwff.rs
@@ -4,7 +4,7 @@ use crate::nn::Initializer;
 use crate::{
     config::Config,
     module::Module,
-    nn::{Dropout, DropoutConfig, Linear, LinearConfig, GELU},
+    nn::{Dropout, DropoutConfig, Gelu, Linear, LinearConfig},
     tensor::{backend::Backend, Tensor},
 };
 
@@ -36,7 +36,7 @@ pub struct PositionWiseFeedForward<B: Backend> {
     linear_inner: Linear<B>,
     linear_outer: Linear<B>,
     dropout: Dropout,
-    gelu: GELU,
+    gelu: Gelu,
 }
 
 impl PositionWiseFeedForwardConfig {
@@ -50,7 +50,7 @@ impl PositionWiseFeedForwardConfig {
                 .with_initializer(self.initializer.clone())
                 .init(device),
             dropout: DropoutConfig::new(self.dropout).init(),
-            gelu: GELU::new(),
+            gelu: Gelu::new(),
         }
     }
     /// Initialize a new [position-wise feed-forward](PositionWiseFeedForward) module with a
@@ -63,7 +63,7 @@ impl PositionWiseFeedForwardConfig {
             linear_inner: LinearConfig::new(self.d_model, self.d_ff).init_with(record.linear_inner),
             linear_outer: LinearConfig::new(self.d_ff, self.d_model).init_with(record.linear_outer),
             dropout: DropoutConfig::new(self.dropout).init(),
-            gelu: GELU::new(),
+            gelu: Gelu::new(),
         }
     }
 }

diff --git a/burn-no-std-tests/src/conv.rs b/burn-no-std-tests/src/conv.rs
@@ -11,7 +11,7 @@ use burn::{
 pub struct ConvBlock<B: Backend> {
     conv: nn::conv::Conv2d<B>,
     pool: nn::pool::MaxPool2d,
-    activation: nn::GELU,
+    activation: nn::Gelu,
 }
 
 #[derive(Config)]
@@ -29,7 +29,7 @@ impl<B: Backend> ConvBlock<B> {
         let pool = nn::pool::MaxPool2dConfig::new(config.kernel_size)
             .with_padding(nn::PaddingConfig2d::Same)
             .init();
-        let activation = nn::GELU::new();
+        let activation = nn::Gelu::new();
 
         Self {
             conv,

diff --git a/examples/mnist-inference-web/README.md b/examples/mnist-inference-web/README.md
@@ -44,10 +44,10 @@ values.
 Layers:
 
 1. Input Image (28,28, 1ch)
-2. `Conv2d`(3x3, 8ch), `BatchNorm2d`, `GELU`
-3. `Conv2d`(3x3, 16ch), `BatchNorm2d`, `GELU`
-4. `Conv2d`(3x3, 24ch), `BatchNorm2d`, `GELU`
-5. `Linear`(11616, 32), `GELU`
+2. `Conv2d`(3x3, 8ch), `BatchNorm2d`, `Gelu`
+3. `Conv2d`(3x3, 16ch), `BatchNorm2d`, `Gelu`
+4. `Conv2d`(3x3, 24ch), `BatchNorm2d`, `Gelu`
+5. `Linear`(11616, 32), `Gelu`
 6. `Linear`(32, 10)
 7. Softmax Output
 

diff --git a/examples/mnist-inference-web/src/model.rs b/examples/mnist-inference-web/src/model.rs
@@ -16,7 +16,7 @@ pub struct Model<B: Backend> {
     dropout: nn::Dropout,
     fc1: nn::Linear<B>,
     fc2: nn::Linear<B>,
-    activation: nn::GELU,
+    activation: nn::Gelu,
 }
 
 const NUM_CLASSES: usize = 10;
@@ -43,7 +43,7 @@ impl<B: Backend> Model<B> {
             fc1,
             fc2,
             dropout,
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
         }
     }
 
@@ -70,7 +70,7 @@ impl<B: Backend> Model<B> {
 pub struct ConvBlock<B: Backend> {
     conv: nn::conv::Conv2d<B>,
     norm: BatchNorm<B, 2>,
-    activation: nn::GELU,
+    activation: nn::Gelu,
 }
 
 impl<B: Backend> ConvBlock<B> {
@@ -83,7 +83,7 @@ impl<B: Backend> ConvBlock<B> {
         Self {
             conv,
             norm,
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
         }
     }
 

diff --git a/examples/mnist/src/model.rs b/examples/mnist/src/model.rs
@@ -17,7 +17,7 @@ pub struct Model<B: Backend> {
     dropout: nn::Dropout,
     fc1: nn::Linear<B>,
     fc2: nn::Linear<B>,
-    activation: nn::GELU,
+    activation: nn::Gelu,
 }
 
 impl<B: Backend> Default for Model<B> {
@@ -51,7 +51,7 @@ impl<B: Backend> Model<B> {
             dropout,
             fc1,
             fc2,
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
         }
     }
 
@@ -92,7 +92,7 @@ impl<B: Backend> Model<B> {
 pub struct ConvBlock<B: Backend> {
     conv: nn::conv::Conv2d<B>,
     norm: BatchNorm<B, 2>,
-    activation: nn::GELU,
+    activation: nn::Gelu,
 }
 
 impl<B: Backend> ConvBlock<B> {
@@ -105,7 +105,7 @@ impl<B: Backend> ConvBlock<B> {
         Self {
             conv,
             norm,
-            activation: nn::GELU::new(),
+            activation: nn::Gelu::new(),
         }
     }