Add File Size (MB) column to benchmarks

[glenn-jocher]

Adds a file size column to benchmarks output, i.e.:

Benchmarks complete (97.27s)
                   Format  Size (MB)  mAP@0.5:0.95  Inference time (ms)
0                 PyTorch        3.9        0.2927                50.13
1             TorchScript        7.4        0.2927                57.51
2                    ONNX        7.2        0.2927                34.06
3                OpenVINO        7.5        0.2927                17.43
4                TensorRT        NaN           NaN                  NaN
5                  CoreML        NaN           NaN                  NaN
6   TensorFlow SavedModel        7.3        0.2927                64.48
7     TensorFlow GraphDef        7.3        0.2927                68.57
8         TensorFlow Lite        3.7        0.2933                18.20
9     TensorFlow Edge TPU        NaN           NaN                  NaN
10          TensorFlow.js        NaN           NaN                  NaN

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhancement of benchmarking output with model file size metric.

📊 Key Changes

• Added file_size() function call to include the file size of the model in the benchmark results.
• Adjusted benchmark result data structure to incorporate model file size.
• Updated DataFrame columns to reflect the new structure with model file size information.

🎯 Purpose & Impact

• Purpose: To provide users with more comprehensive information on model performance by including model size alongside accuracy (mAP) and inference time.
• Impact: Users can now consider model size when evaluating model performance, which is particularly important for deploying models on resource-constrained environments. 📉📲