chore: fix some comments (#2717)

Signed-off-by: sunxunle <[email protected]>

Author: sunxunle

crates/burn-import/src/pytorch/recorder.rs

@@ -11,7 +11,7 @@ use serde::{de::DeserializeOwned, Serialize};
 
 use super::reader::from_file;
 
-/// A recorder that that loads PyTorch files (`.pt`) into Burn modules.
+/// A recorder that loads PyTorch files (`.pt`) into Burn modules.
 ///
 /// LoadArgs can be used to remap keys or file path.
 /// See [LoadArgs](struct.LoadArgs.html) for more information.

crates/burn-train/src/checkpoint/strategy/metric.rs

@@ -114,7 +114,7 @@ mod tests {
         process_train(&mut processor, 0.3, epoch);
         end_epoch(&mut processor, epoch);
 
-        // Should save the current record and delete the pervious one.
+        // Should save the current record and delete the previous one.
         assert_eq!(
             vec![CheckpointingAction::Delete(1), CheckpointingAction::Save],
             strategy.checkpointing(epoch, &store)

examples/custom-image-dataset/src/dataset.rs

@@ -5,7 +5,7 @@ use tar::Archive;
 use burn::data::{dataset::vision::ImageFolderDataset, network::downloader};
 
 /// CIFAR-10 mirror from [fastai](https://github.com/fastai/fastai/blob/master/fastai/data/external.py#L44).
-/// Licensed under the [Appache License](https://github.com/fastai/fastai/blob/master/LICENSE).
+/// Licensed under the [Apache License](https://github.com/fastai/fastai/blob/master/LICENSE).
 const URL: &str = "https://s3.amazonaws.com/fast-ai-sample/cifar10.tgz";
 
 /// The [CIFAR-10](https://www.cs.toronto.edu/%7Ekriz/cifar.html) dataset consists of 60,000 32x32