This document describes evaluation of optimized checkpoints for SegNet
Please install and setup AIMET before proceeding further.
This model was tested with the torch_gpu
variant of AIMET 1.23.
pip install scikit-image
- Clone the aimet-model-zoo repo
git clone https://github.com/quic/aimet-model-zoo.git
- Append the repo location to your
PYTHONPATH
with the following:
export PYTHONPATH=$PYTHONPATH:<path to aimet_model_zoo_path>
This evaluation was designed for the CamVid dataset variant provided by SegNet authors repository.
- Download and extract CamVid directory:
wget https://github.com/alexgkendall/SegNet-Tutorial/archive/refs/heads/master.zip
unzip master.zip && mv SegNet-Tutorial-master/CamVid . && rm -r SegNet-Tutorial-master
- The SegNet model checkpoints can be downloaded from the Releases page.
To run evaluation with QuantSim in AIMET, use the following
python3 aimet-model-zoo/aimet_zoo_torch/segnet/evaluator/segnet_quanteval.py \
--dataset-path <path to CamVid dataset> \
--model-config <configuration to be tested>
Available model configurations are:
- segnet_w8a8
- segnet_w4a8
- Weight quantization: 8 or 4 bits, per channel symmetric quantization
- Bias parameters are not quantized
- Activation quantization: 8 bits, asymmetric quantization
- Model inputs are quantized
- TF_enhanced was used as quantization scheme