Fix TFLite path, add confidence histogram placeholder, refine download instructions
Browse files
README.md
CHANGED
|
@@ -24,7 +24,7 @@ widget:
|
|
| 24 |
src: https://cdn-uploads.huggingface.co/production/uploads/66cdac913f233bf2c7b4f590/HzXxNze2jmCkV5KPY_fpQ.png
|
| 25 |
example_title: Bridge Classification
|
| 26 |
model-index:
|
| 27 |
-
- name:
|
| 28 |
results:
|
| 29 |
- task:
|
| 30 |
type: image-classification
|
|
@@ -45,7 +45,7 @@ model-index:
|
|
| 45 |
name: TTA Accuracy
|
| 46 |
---
|
| 47 |
|
| 48 |
-
#
|
| 49 |
|
| 50 |
Fine-Grained Visual Categorization (FGVC) of world architectural buildings using CNN transfer learning with DenseNet121, enhanced with GeM Pooling, Focal Loss, and Stochastic Weight Averaging (SWA).
|
| 51 |
|
|
@@ -402,7 +402,7 @@ from PIL import Image
|
|
| 402 |
|
| 403 |
LABELS = ["bridge", "castle", "mosque", "skyscraper", "stadium", "temple"]
|
| 404 |
|
| 405 |
-
interpreter = tf.lite.Interpreter(model_path="model.tflite")
|
| 406 |
interpreter.allocate_tensors()
|
| 407 |
|
| 408 |
input_details = interpreter.get_input_details()
|
|
@@ -503,7 +503,7 @@ print(f"Predicted: {LABELS[np.argmax(preds)]} ({np.max(preds)*100:.1f}%)")
|
|
| 503 |
|
| 504 |
```bibtex
|
| 505 |
@misc{saugani2026_arch_building,
|
| 506 |
-
title={Fine-Grained Visual Categorization of World Architectural Buildings
|
| 507 |
Using CNN Transfer Learning DenseNet121 with Fine-Tuning and
|
| 508 |
Multi-Layer Regularization Strategy},
|
| 509 |
author={Saugani},
|
|
|
|
| 24 |
src: https://cdn-uploads.huggingface.co/production/uploads/66cdac913f233bf2c7b4f590/HzXxNze2jmCkV5KPY_fpQ.png
|
| 25 |
example_title: Bridge Classification
|
| 26 |
model-index:
|
| 27 |
+
- name: Architectural Building Image Classifier
|
| 28 |
results:
|
| 29 |
- task:
|
| 30 |
type: image-classification
|
|
|
|
| 45 |
name: TTA Accuracy
|
| 46 |
---
|
| 47 |
|
| 48 |
+
# Architectural Building Image Classifier
|
| 49 |
|
| 50 |
Fine-Grained Visual Categorization (FGVC) of world architectural buildings using CNN transfer learning with DenseNet121, enhanced with GeM Pooling, Focal Loss, and Stochastic Weight Averaging (SWA).
|
| 51 |
|
|
|
|
| 402 |
|
| 403 |
LABELS = ["bridge", "castle", "mosque", "skyscraper", "stadium", "temple"]
|
| 404 |
|
| 405 |
+
interpreter = tf.lite.Interpreter(model_path="tflite/model.tflite")
|
| 406 |
interpreter.allocate_tensors()
|
| 407 |
|
| 408 |
input_details = interpreter.get_input_details()
|
|
|
|
| 503 |
|
| 504 |
```bibtex
|
| 505 |
@misc{saugani2026_arch_building,
|
| 506 |
+
title={Fine-Grained Visual Categorization (FGVC) of World Architectural Buildings
|
| 507 |
Using CNN Transfer Learning DenseNet121 with Fine-Tuning and
|
| 508 |
Multi-Layer Regularization Strategy},
|
| 509 |
author={Saugani},
|