Instructions below :)

Create Conda environment first

conda create python=3.10 -n benchmark_qai_hub

Activate environment

conda activate benchmark_qai_hub

Install requirements

pip install -r  .\requirements.txt

Run GUI

python. .\benchmark_gui.py

Notes

Preferably your paths should look like this

📂 BenchmarkQAIHub/
├─── 📁 datasets_onnx
├─── 📁 datasets_quantized_onnx
├─── 📁 datasets_quantized_tflite
├─── 📁 datasets_tflite
├─── 🐍 benchmark_gui.py
├─── 🐍 one_script_to_rule_them_all.py
├─── 📃 class_index.json
├─── 📃 ground_truth.json
├─── 📃 synset.json
├─── 📄 requirements.txt
├─── 📄 model_accuracy_scores.txt
└─── etc.

To Do

• Keep an individual log of inference fails
• Add button to recalculate accuracy given results directory
• Change Datasets Directory Path entry to accept full paths
• [ ]