| language: en | |
| license: mit | |
| tags: | |
| - semantic-segmentation | |
| - pytorch | |
| - deeplabv3plus | |
| - resnet34 | |
| - agricultural-cv | |
| model-index: | |
| - name: nematode-deeplabv3plus_512-run_2026-06-28_17-57_sample_50 | |
| results: | |
| - task: | |
| type: semantic-segmentation | |
| name: Corm Segmentation | |
| metrics: | |
| - type: mean_iou | |
| value: 0.8735190836786705 | |
| name: Best Val Mean IoU | |
| - type: damage_iou | |
| value: 0.7381499204526462 | |
| name: Best Val Damage-Class IoU | |
| - type: root_iou | |
| value: 0.8856007365241481 | |
| name: Best Val Root-Class IoU | |
| # DeepLabV3+ Corm & Damage Semantic Segmentation Model (Ablation: 70.0%) | |
| This repository contains weights, performance curves, and hyperparameter logs for experiment `nematode-deeplabv3plus_512-run_2026-06-28_17-57_sample_50`. | |
| ## Model Summary | |
| - **Architecture Type:** DeepLabV3+ (SMP wrapper) | |
| - **Backbone Encoder:** `resnet34` | |
| - **Training Resolution:** `512x512` | |
| - **Loss Optimizer configuration:** `CrossEntropy` | |
| - **Dataset Sample Fraction:** `70.0%` | |
| ## Metrics Curves | |
| Below are the training performance plots generated for this run: | |
|  |