Format.

Author: twitchax

src/bin.rs

@@ -463,7 +463,9 @@ fn start(args: Args) -> Void {
                         klib::ml::train::run_training::<Autodiff<NdArray<f32>>>(device, &config, true, true)?;
                     }
                     _ => {
-                        return Err(anyhow::Error::msg("Invalid device (must choose either `gpu` [requires `ml_gpu` feature], `wgpu` [requires `ml_gpu` feature] or `cpu`)."));
+                        return Err(anyhow::Error::msg(
+                            "Invalid device (must choose either `gpu` [requires `ml_gpu` feature], `wgpu` [requires `ml_gpu` feature] or `cpu`).",
+                        ));
                     }
                 }
             }

src/core/pitch.rs

@@ -32,15 +32,15 @@ pub trait HasFrequency {
     /// Essentially, mid way between the frequency and the next frequency on either side.
     fn frequency_range(&self) -> (f32, f32) {
         let frequency = self.frequency();
-        
+
         (frequency * (1.0 - 1.0 / 17.462 / 2.0), frequency * (1.0 + 1.0 / 16.8196 / 2.0))
     }
 
     /// Returns the tight frequency range of the type (usually a [`Note`]).
     /// Essentially, 1/8 the way between the frequency and the next frequency on either side.
     fn tight_frequency_range(&self) -> (f32, f32) {
         let frequency = self.frequency();
-        
+
         (frequency * (1.0 - 1.0 / 17.462 / 8.0), frequency * (1.0 + 1.0 / 16.8196 / 8.0))
     }
 }

src/ml/base/data.rs

@@ -3,7 +3,7 @@
 use burn::tensor::{backend::Backend, Data, Tensor};
 
 use super::{
-    helpers::{get_deterministic_guess, mel_filter_banks_from, u128_to_binary, note_binned_convolution},
+    helpers::{get_deterministic_guess, mel_filter_banks_from, note_binned_convolution, u128_to_binary},
     KordItem, INPUT_SPACE_SIZE, NUM_CLASSES,
 };
 

src/ml/base/gather.rs

@@ -2,14 +2,14 @@
 
 use std::path::Path;
 
+use crate::core::{
+    base::{Parsable, Void},
+    note::{HasNoteId, Note},
+};
 use crate::{
     analyze::base::{get_frequency_space, get_smoothed_frequency_space},
     ml::base::{KordItem, FREQUENCY_SPACE_SIZE},
 };
-use crate::core::{
-        base::{Parsable, Void},
-        note::{HasNoteId, Note},
-    };
 
 use crate::analyze::mic::get_audio_data_from_microphone;
 

src/ml/base/helpers.rs

@@ -8,15 +8,19 @@ use std::{
     path::{Path, PathBuf},
 };
 
-use burn::{tensor::{backend::Backend, Tensor}, module::Module};
+use burn::{
+    module::Module,
+    tensor::{backend::Backend, Tensor},
+};
 use byteorder::{BigEndian, ReadBytesExt, WriteBytesExt};
 
 use crate::{
     analyze::base::get_notes_from_smoothed_frequency_space,
     core::{
         base::Res,
         helpers::{inv_mel, mel},
-        note::{HasNoteId, Note, ALL_PITCH_NOTES_WITH_FREQUENCY}, pitch::HasFrequency,
+        note::{HasNoteId, Note, ALL_PITCH_NOTES_WITH_FREQUENCY},
+        pitch::HasFrequency,
     },
 };
 
@@ -113,7 +117,7 @@ pub fn note_binned_convolution(spectrum: &[f32]) -> [f32; NUM_CLASSES] {
 
     for (note, _) in ALL_PITCH_NOTES_WITH_FREQUENCY.iter().skip(7).take(90) {
         let id_index = note.id_index();
-        
+
         let (low, high) = note.tight_frequency_range();
         let low = low.round() as usize;
         let high = high.round() as usize;
@@ -137,13 +141,7 @@ pub fn note_binned_convolution(spectrum: &[f32]) -> [f32; NUM_CLASSES] {
 pub fn harmonic_convolution(spectrum: &[f32]) -> [f32; FREQUENCY_SPACE_SIZE] {
     let mut harmonic_convolution = [0f32; FREQUENCY_SPACE_SIZE];
 
-    let (peak, _) = spectrum.iter().enumerate().fold((0usize, 0f32), |(k, max), (j, x)| {
-        if *x > max {
-            (j, *x)
-        } else {
-            (k, max)
-        }
-    });
+    let (peak, _) = spectrum.iter().enumerate().fold((0usize, 0f32), |(k, max), (j, x)| if *x > max { (j, *x) } else { (k, max) });
 
     for center in (peak / 2)..4000 {
         let mut sum = spectrum[center];
@@ -239,4 +237,4 @@ impl<B: Backend> Sigmoid<B> {
         //let scaled = input;
         scaled.clone().exp().div(scaled.exp().add_scalar(1.0))
     }
-}
+}

src/ml/base/mlp.rs

@@ -1,8 +1,9 @@
 //! Multilayer Perceptron module.
 
 use burn::{
-    nn::{self, LayerNormConfig, LayerNorm},
-    tensor::{backend::Backend, Tensor}, module::Module,
+    module::Module,
+    nn::{self, LayerNorm, LayerNormConfig},
+    tensor::{backend::Backend, Tensor},
 };
 
 /// Multilayer Perceptron module.
@@ -28,12 +29,7 @@ impl<B: Backend> Mlp<B> {
         let dropout = nn::DropoutConfig::new(mlp_dropout).init();
         let activation = nn::ReLU::new();
 
-        Self {
-            linears,
-            norm,
-            dropout,
-            activation
-        }
+        Self { linears, norm, dropout, activation }
     }
 
     /// Applies the forward pass on the input tensor.

src/ml/base/model.rs

@@ -3,20 +3,21 @@
 use core::f32;
 
 use burn::{
-    nn::{self, attention::{MultiHeadAttentionConfig, MultiHeadAttention, MhaInput}},
-    tensor::{backend::Backend, Tensor}, module::Module,
+    module::Module,
+    nn::{
+        self,
+        attention::{MhaInput, MultiHeadAttention, MultiHeadAttentionConfig},
+    },
+    tensor::{backend::Backend, Tensor},
 };
 
 use super::{helpers::Sigmoid, INPUT_SPACE_SIZE, NUM_CLASSES};
 
 #[cfg(feature = "ml_train")]
-use crate::ml::train::{
-    data::KordBatch,
-    helpers::KordClassificationOutput,
-};
+use crate::ml::train::{data::KordBatch, helpers::KordClassificationOutput};
 
 /// The Kord model.
-/// 
+///
 /// This model is a transformer model that uses multi-head attention to classify notes from a frequency space.
 #[derive(Module, Debug)]
 pub struct KordModel<B: Backend> {
@@ -32,11 +33,7 @@ impl<B: Backend> KordModel<B> {
         let output = nn::LinearConfig::new(INPUT_SPACE_SIZE, NUM_CLASSES).init::<B>();
         let sigmoid = Sigmoid::new(sigmoid_strength);
 
-        Self {
-            mha,
-            output,
-            sigmoid,
-        }
+        Self { mha, output, sigmoid }
     }
 
     /// Applies the forward pass on the input tensor.
@@ -50,14 +47,14 @@ impl<B: Backend> KordModel<B> {
 
         // Reshape the output to remove the sequence dimension.
         let mut x = attn.context.reshape([batch_size, input_size]);
-        
+
         // Perform the final linear layer to map to the output dimensions.
         x = self.output.forward(x);
 
         // Apply the sigmoid function to the output to achieve multi-classification.
         x = self.sigmoid.forward(x);
 
-        x 
+        x
     }
 
     /// Applies the forward classification pass on the input tensor.
@@ -90,4 +87,4 @@ impl<B: Backend> KordModel<B> {
 
         KordClassificationOutput { loss, output, targets }
     }
-}
+}

src/ml/infer/execute.rs

@@ -3,7 +3,8 @@
 use burn::{
     config::Config,
     module::Module,
-    tensor::backend::Backend, record::{BinBytesRecorder, FullPrecisionSettings, Recorder},
+    record::{BinBytesRecorder, FullPrecisionSettings, Recorder},
+    tensor::backend::Backend,
 };
 use burn_ndarray::{NdArray, NdArrayDevice};
 use serde::{de::DeserializeOwned, Serialize};

src/ml/train/execute.rs

@@ -3,13 +3,15 @@
 use std::sync::Arc;
 
 use burn::{
+    backend::Autodiff,
     config::Config,
     data::dataloader::DataLoaderBuilder,
+    lr_scheduler::constant::ConstantLr,
     module::Module,
     optim::{decay::WeightDecayConfig, AdamConfig},
-    backend::Autodiff,
+    record::{BinFileRecorder, FullPrecisionSettings, Recorder},
     tensor::backend::{AutodiffBackend, Backend},
-    train::{metric::LossMetric, LearnerBuilder}, lr_scheduler::constant::ConstantLr, record::{BinFileRecorder, FullPrecisionSettings, Recorder},
+    train::{metric::LossMetric, LearnerBuilder},
 };
 use serde::{de::DeserializeOwned, Serialize};
 
@@ -31,8 +33,8 @@ use super::{
 use crate::ml::base::TrainConfig;
 
 /// Run the training.
-/// 
-/// Given the [`TrainConfig`], this function will run the training and return the overall accuracy on 
+///
+/// Given the [`TrainConfig`], this function will run the training and return the overall accuracy on
 /// the validation / test set.
 pub fn run_training<B: AutodiffBackend>(device: B::Device, config: &TrainConfig, print_accuracy_report: bool, save_model: bool) -> Res<f32>
 where
@@ -157,7 +159,7 @@ pub fn compute_overall_accuracy<B: Backend>(model_trained: &KordModel<B>, device
 }
 
 /// Run hyper parameter tuning.
-/// 
+///
 ///This method sweeps through the hyper parameters and runs training for each combination. The best
 /// hyper parameters are then printed at the end.
 #[coverage(off)]

src/ml/train/helpers.rs

@@ -8,9 +8,8 @@ use burn::{
     },
     train::{
         metric::{
-            MetricMetadata,
             state::{FormatOptions, NumericMetricState},
-            Adaptor, LossInput, Metric, MetricEntry, Numeric,
+            Adaptor, LossInput, Metric, MetricEntry, MetricMetadata, Numeric,
         },
         TrainOutput, TrainStep, ValidStep,
     },
@@ -24,11 +23,7 @@ use crate::{
         note::{HasNoteId, Note, ALL_PITCH_NOTES},
         pitch::HasFrequency,
     },
-    ml::base::{
-        helpers::load_kord_item,
-        model::KordModel,
-        KordItem, FREQUENCY_SPACE_SIZE, NUM_CLASSES,
-    },
+    ml::base::{helpers::load_kord_item, model::KordModel, KordItem, FREQUENCY_SPACE_SIZE, NUM_CLASSES},
 };
 
 use super::data::KordBatch;