Please install and setup AIMET before proceeding further. This evaluation was run using AIMET 1.22.2 for TensorFlow 1.15 i.e. please set release_tag="1.22.2" and AIMET_VARIANT="tf_gpu_tf115" in the above instructions.
This model requires the following python package versions:
pip install tensorflow-gpu==1.15.0
-
Clone the TensorFlow Models repo
git clone https://github.com/tensorflow/models.git
cd models -
Checkout this commit id:
git checkout 104488e40bc2e60114ec0212e4e763b08015ef97 -
Append the repo location to your
PYTHONPATHwith the following:export PYTHONPATH=$PYTHONPATH:<path to tensorflow models repo>/research/slim
export PYTHONPATH=$PYTHONPATH:<path to parent>/aimet-model-zoo
Note: This model is expected not to work with GPUs at or after NVIDIA 30-series (e.g. RTX 3050), as those bring a new architecture not fully compatible with TF 1.X
- ImageNet can be downloaded from here:
The dataset must then be converted to TFRecords formatting. The resulting directory should contain TFRecords files named as: train-00000-of-01024 up to train-01023-of-01024 and validation-00001-of-00128 up to validation-00127-of-00128. This evalution script will only use validation data.
- The original ResNet 50 checkpoint is available on the TensorFlow Models repo.
python resnet50_v1_quanteval.py \
--dataset-path <path to imagenet validation TFRecords> \
--eval_quantized <True evaluates the optimized model, False the original model>Setting eval_quantized=True will evaluate the optimized model's performance on both GPU and on a simulated hardware device.
Similarly, eval_quantized=False will evalute the original source model on both GPU and simulated device.
- Weight quantization: 8 bits, per tensor asymmetric quantization
- Bias parameters are quantized
- Activation quantization: 8 bits, asymmetric quantization
- Operations which shuffle data such as reshape or transpose do not require additional quantizers