OncoVision-X / src /models /baseline_resnet3d.py
adityasync's picture
Clean OncoVision-X deployment with LFS
8960670
raw
history blame contribute delete
756 Bytes
#!/usr/bin/env python3
"""
Baseline 3D ResNet-18 — Standalone module.
This module provides the ResNet3D18 class as a standalone import
per the ablation study specification. The actual implementation
lives in baselines.py; this file re-exports it for clean per-model imports.
Usage:
from src.models.baseline_resnet3d import ResNet3D18
model = ResNet3D18(num_classes=1)
Architecture:
- Backbone: torchvision video.r3d_18 (3D ResNet-18)
- Input conv modified: 3-channel → 1-channel (grayscale CT)
- Output: binary logit (B, 1)
- Forward signature: forward(nodule_patch, context_patch=None)
(context_patch is ignored — baseline uses nodule only)
"""
from src.models.baselines import ResNet3D18
__all__ = ['ResNet3D18']