Table 5 Hardware and software specifications.
Component | Workstation | Laptop |
|---|---|---|
Processor | Intel Xeon W-3275 (28 cores, 2.5GHz base, 4.4GHz boost) | Intel Core i9-12900H (14 cores, 2.5GHz base, 5.0GHz boost) |
RAM | 256GB DDR4-3200 ECC | 64GB DDR5-4800 |
GPU primary | NVIDIA RTX A6000 (48GB GDDR6) | NVIDIA RTX 4090 Mobile (16GB GDDR6) |
GPU secondary | NVIDIA RTX A5000 (24GB GDDR6) | N/A |
Storage | 4TB NVMe SSD RAID 0 + 24TB HDD (RAID 5) | 2TB NVMe SSD |
Operating system | Ubuntu 22.04 LTS | Ubuntu 22.04 LTS |
CUDA version | 12.2 | 12.1 |
cuDNN version | 8.9.4 | 8.9.2 |
Python version | 3.10.12 | 3.10.12 |
Deep learning framework | PyTorch 2.1.0 | PyTorch 2.0.1 |
Image processing | OpenCV 4.8.0, scikit-image 0.21.0 | OpenCV 4.7.0, scikit-image 0.20.0 |
Data management | pandas 2.1.1, NumPy 1.25.2 | pandas 2.0.3, NumPy 1.24.3 |
Visualization | Matplotlib 3.7.2, Tensorboard 2.13.0 | Matplotlib 3.7.1, Tensorboard 2.12.3 |
Optimization libraries | NVIDIA DALI, TorchVision 0.16.0 | TorchVision 0.15.2 |
Memory utilization | Peak: 44GB GPU, 212GB RAM | Peak: 14GB GPU, 48GB RAM |
Training duration | ~ 72 h (full dataset) | ~ 12 h (subset testing) |
Inference speed | 0.37s per image | 1.24s per image |
