Upload README.md with huggingface_hub

README.md

@@ -9,18 +9,71 @@ tags:
 - insect-detection
 - biology
 - agriculture
+- fine-tuned
 license: mit
 pipeline_tag: image-classification
+metrics:
+- accuracy
+- f1
+model-index:
+- name: Mustafa5645344/insect-detection-vit
+  results:
+  - task:
+      type: image-classification
+      name: Insect Species Classification
+    dataset:
+      type: specifly-insect-dataset
+      name: Specifly 50K Insect Dataset
+    metrics:
+    - type: accuracy
+      value: 0.9851
+      name: Accuracy
+    - type: f1
+      value: 0.9858
+      name: F1-Score
 ---
 
 # Insect Classification: Vision Transformer (ViT) Model
 
-This repository contains a fine-tuned **Vision Transformer (ViT)** model (`ViTForImageClassification`) trained to classify **21 distinct insect species** with high precision for agricultural monitoring and biological biodiversity analysis.
+This model is a high-accuracy fine-tuned **Vision Transformer (ViT)** (`google/vit-base-patch16-224-in21k` backbone) trained to classify **21 distinct agricultural insect pest and beneficial species**.
 
-## Supported Insect Classes (21 Classes)
-- `ant` (0), `aphid` (1), `bees` (2), `butterfly` (3), `caterpillar` (4), `cicada` (5), `dragonfly` (6), `grasshopper` (7), `green_lacewing` (8), `ladybug` (9), `leafhopper` (10), `mantis` (11), `mole_cricket` (12), `planthopper` (13), `rhino_beetle` (14), `rice_bug` (15), `spider` (16), `stem_borer` (17), `stink_bug` (18), `undefined` (19), `weevil` (20)
+## Training Hardware & Specs
+- **Compute:** NVIDIA A100 Tensor Core GPU (SXM4 40GB)
+- **Total Training Steps:** 16,600 iterations (20 full epochs)
+- **Training Time:** 31 minutes 25 seconds
+- **Base Architecture:** Vision Transformer Base (`vit-base-patch16-224`)
+- **Image Resolution:** 224 x 224 px
 
-## Quick Usage
+---
+
+## Evaluation Benchmark & Training History
+
+| Epoch | Training Loss | Validation Loss | Accuracy | F1-Score |
+| :---: | :---: | :---: | :---: | :---: |
+| 1 | 0.3389 | 0.3206 | 97.41% | 97.57% |
+| 2 | 0.1076 | 0.1481 | 97.77% | 97.91% |
+| 3 | 0.0486 | 0.1073 | 97.83% | 97.95% |
+| 4 | 0.0252 | 0.0999 | 98.13% | 98.18% |
+| 5 | 0.0167 | 0.0784 | 98.45% | 98.51% |
+| 10 | 0.0065 | 0.1190 | 98.03% | 98.12% |
+| 15 | 0.0003 | 0.1093 | 98.42% | 98.50% |
+| 19 | 0.0001 | 0.1151 | **98.51%** | **98.59%** |
+| **20 (Final)** | **0.0001** | **0.1153** | **98.51%** | **98.58%** |
+
+### Top Metrics Achieved:
+- **Final Top-1 Accuracy:** **98.51%**
+- **Macro F1-Score:** **98.58%**
+- **Minimal Training Loss:** **0.000100**
+
+---
+
+## Supported Taxa / Classes (21 Classes)
+
+`ant`, `aphid`, `bees`, `butterfly`, `caterpillar`, `cicada`, `dragonfly`, `grasshopper`, `green_lacewing`, `ladybug`, `leafhopper`, `mantis`, `mole_cricket`, `planthopper`, `rhino_beetle`, `rice_bug`, `spider`, `stem_borer`, `stink_bug`, `undefined`, `weevil`
+
+---
+
+## Quick Inference Usage
 
 ```python
 from transformers import AutoImageProcessor, AutoModelForImageClassification
@@ -38,8 +91,11 @@ inputs = processor(images=image, return_tensors="pt")
 
 with torch.no_grad():
     outputs = model(**inputs)
-    logits = outputs.logits
-    predicted_class_idx = logits.argmax(-1).item()
+    probs = torch.softmax(outputs.logits, dim=-1)
+    predicted_idx = outputs.logits.argmax(-1).item()
+
+predicted_label = model.config.id2label[str(predicted_idx)]
+confidence = probs[0][predicted_idx].item()
 
-print("Predicted Insect:", model.config.id2label[str(predicted_class_idx)])
+print(f"Species: {predicted_label} (Confidence: {confidence*100:.2f}%)")
 ```