dmetan/classification_servers_cnn_resnet50 repository overview

This Docker image provides a deep learning project for server equipment recognition using a fine‑tuned ResNet50 CNN. The model is trained to classify hardware into five categories: commutators, firewalls, routers, servers, switches.

Technical Details Framework: PyTorch 2.5.1 + CUDA 12.1 + cuDNN9

Model: ResNet50, fine‑tuned for hardware classification

Training setup: GPU acceleration (tested on NVIDIA RTX 3090), AMP mixed precision, Early Stopping, logging, progress bars

Dataset: ~90 images split into training and validation sets

Metrics: Accuracy, Precision, Recall, F1‑score, Confusion Matrix

Visualization: loss/accuracy plots, confusion matrix, sample predictions

🚀 Quick Start Pull the image:

docker pull dmetan/classification_servers_cnn_resnet50:latest

Run the container:

docker run --rm -it --gpus all dmetan/classification_servers_cnn_resnet50:latest