ImageFolderDatasetGuide Modification

docs/imagefolder-guide.md

@@ -1,6 +1,10 @@
 # ImageFolderDataset Guide: Working with NEON Beetle Images
 
-This guide demonstrates how to use the [`ImageFolderDataset`](api-reference.md#hpc_inference.datasets.ImageFolderDataset) class from the HPC Inference package with the [2018 NEON Beetles dataset](https://huggingface.co/datasets/imageomics/2018-NEON-beetles).
+This guide demonstrates how to use the [`ImageFolderDataset`](api-reference.md#hpc_inference.datasets.ImageFolderDataset) class from the `hpc-inference` package with the 2018 NEON Beetles dataset.[^1]
+
+[^1]: Fluck, I. E., Baiser, B., Wolcheski, R., Chinmian, I., & Record, S. (2024). 
+2018 NEON Ethanol-preserved Ground Beetles (Revision 7b3731d). 
+Hugging Face. https://doi.org/10.57967/hf/5252
 
 ## Overview
 
@@ -14,7 +18,7 @@ The [`ImageFolderDataset`](api-reference.md#hpc_inference.datasets.ImageFolderDa
 
 ## Dataset Overview
 
-The 2018 NEON Beetles dataset contains:
+The 2018 NEON Beetles dataset[^1] contains:
 - 577 high-resolution group images of ethanol-preserved beetles
 - Beetles collected from NEON ecological sites in 2018
 - Multiple beetles per image, organized by species
@@ -458,7 +462,8 @@ plt.show()
 ```
 ![Original vs Processed Beetle Images](imgs/processed_beetles.png)
 
-The top row shows the original high-resolution images (5568×3712 pixels). The bottom row shows the same images after ResNet preprocessing: resized to 224×224 pixels, center-cropped, and de-normalized.
+*Figure 1: Comparison of original vs processed NEON beetle images from the 2018 NEON Beetles dataset[^1]. Top row shows original high-resolution images (5568×3712 pixels), bottom row shows the same images after ResNet preprocessing: resized to 224×224 pixels, center-cropped, and normalized.*
+
 
 ## Performance Optimization
 
@@ -572,4 +577,5 @@ Based on the examples above, here are key recommendations for using [`ImageFolde
 ### Collate Functions
 - Use [`pil_image_collate`](api-reference.md#hpc_inference.utils) when `preprocess=None` (returning PIL Images)
 - Use [`multi_model_collate`](api-reference.md#hpc_inference.utils) for multi-model preprocessing
-- Use default collate function for single tensor outputs
+- Use default collate function for single tensor outputs
+

docs/stylesheets/extra.css

@@ -0,0 +1,25 @@
+.md-footer-social {
+  order: 1;
+}
+
+.md-copyright {
+  order: 2;
+}
+
+.md-footer-generator {
+  order: 3;
+}
+
+/*
+Make logo larger to be legible,
+based on this suggestion: https://github.com/squidfunk/mkdocs-material/discussions/2933#discussioncomment-1168075
+*/
+ .md-header__button.md-logo {
+  margin: 0;
+  padding: 1;
+}
+
+.md-header__button.md-logo img, .md-header__button.md-logo svg {
+  height: 2.6rem;
+  width: 2.6rem;
+}

mkdocs.yml

@@ -15,8 +15,8 @@ nav:
 
 theme:
   name: material
-  logo: imgs/Imageomics_logo_butterfly.png
-  favicon: imgs/Imageomics_logo_butterfly.png
+  logo: imgs/Imageomics_ABC.png
+  favicon: imgs/Imageomics_ABC.png
   palette:
     # Palette toggle for automatic mode
     - media: "(prefers-color-scheme)"
@@ -46,6 +46,7 @@ theme:
     - search.highlight
 
 markdown_extensions:
+  - footnotes
   - admonition
   - pymdownx.details
   - pymdownx.superfences
@@ -69,6 +70,9 @@ plugins:
         python:
           paths: [src]
 
+extra_css:
+  - stylesheets/extra.css
+
 extra:
   social:
     - icon: fontawesome/brands/github