Update README.md

README.md

@@ -50,14 +50,12 @@ The finetuned models available in this repository are based on following archite
 - [MobileNetV3](https://huggingface.co/timm/mobilenetv3_small_100.lamb_in1k) (small and large version)
 - [MobileViT v2](https://huggingface.co/apple/mobilevitv2-1.0-imagenet1k-256)
 
-| Architecture | Validation Accuracy | Test Accuracy | F1 | Precision | Recall | AUC |
-|--------------|--------------------:|--------------:|---:|----------:|-------:|----:|
-| MobileNetV2 | 1.0000 | 0.9776 | 0.9770 | 1.0000 | 0.9550 | 0.9993 |
-| MobileNetV3 Small | 1.0000 | 0.9731 | 0.9727 | 0.9817 | 0.9640 | 0.9984 |
-| MobileNetV3 Large | 1.0000 | 0.9865 | 0.9864 | 0.9909 | 0.9820 | 0.9992 |
-| MobileViT v2 | 0.9955 | 0.9776 | 0.9770 | 1.0000 | 0.9550 | 0.9992 |
-
-> This repository contains two model variants: models prefix with "final_" are finetuned with no test set
+| Architecture | Validation Accuracy | Test Accuracy |
+|--------------|--------------------:|--------------:|
+| MobileNetV2 | 0.995 | 0.982 |
+| MobileNetV3 Small | 0.991 | 0.968 |
+| MobileNetV3 Large | 1.0 | 0.986 |
+| MobileViT v2 | 0.995 | 0.977 |
 
 > These results should be interpreted with care. Although the models reach very high scores on the current splits, the task may be partially dataset-dependent.
 
@@ -113,9 +111,12 @@ Examples include:
 
 | Illuminated | Not Illuminated |
 |:-----------:|:---------------:|
-| ![illuminated](illuminated_0.jpg) | ![not illuminated](not_illuminated_0.jpg) |
-|  |  |
-
+| <img src="assets/illuminated/1.png" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/1.jpeg" width="150" height="150" style="object-fit:cover;"> |
+| <img src="assets/illuminated/2.jpg" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/2.jpg" width="150" height="150" style="object-fit:cover;"> |
+| <img src="assets/illuminated/3.png" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/3.jpg" width="150" height="150" style="object-fit:cover;"> |
+| <img src="assets/illuminated/4.jpg" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/4.jpg" width="150" height="150" style="object-fit:cover;"> |
+| <img src="assets/illuminated/5.jpg" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/5.png" width="150" height="150" style="object-fit:cover;"> |
+| <img src="assets/illuminated/6.jpg" width="150" height="150" style="object-fit:cover;"> | <img src="assets/not_illuminated/6.jpg" width="150" height="150" style="object-fit:cover;"> |
 # Usage
 
 ## Python — ONNX local
@@ -356,11 +357,17 @@ If you use these models in your research, please cite:
 
 # Funding
 
-<div style="display: flex; align-items: center; justify-content: center; text-align: justify; gap: 20px; max-width: 800px; margin: auto;">
-    <img src="assets/odil-logo.png" width="200" alt="Logo ODIL" align="left">
-  <p style="text-align: justify; margin-top:-20px;">
-  <br>
-This models are developped at [École nationale des chartes – PSL](https://www.chartes.psl.eu/) in context of [O.D.I.L. project](https://projet.biblissima.fr/fr/appels-projets/projets-retenus/odil-objet-detection-illuminations).
-    </p>
-    <br>
+<div style="display: flex; align-items: center; justify-content: center; gap: 20px; max-width: 800px; margin: 0 auto;">
+  <img src="./assets/odil-logo.png" width="90" alt="Logo ODIL" style="flex: 0 0 auto;">
+
+  <p style="text-align: justify; margin: 0;">
+    These models were developed at
+    <a href="https://www.chartes.psl.eu/" target="_blank" rel="noopener">
+      École nationale des chartes – PSL
+    </a>
+    in the context of the
+    <a href="https://projet.biblissima.fr/fr/appels-projets/projets-retenus/odil-objet-detection-illuminations" target="_blank" rel="noopener">
+      O.D.I.L. project
+    </a>.
+  </p>
 </div>