Skip to content

Tutorial_V2X-Real_PTQ.md

Latest commit

 

History

History
18 lines (12 loc) · 1.2 KB

Tutorial_V2X-Real_PTQ.md

File metadata and controls

18 lines (12 loc) · 1.2 KB

Tutorial on Post-Training Quantization (PTQ) Stage on V2X-Real dataset

Please refer to Tutorial of Baseline Training and Inference on V2X-Real dataset and Tutorial of Codebook Learning on V2X-Real dataset before reading this documentation. Post-Training quantization (PTQ) involves the curation of calibration dataset and the calibration process as described in the paper.

Post-Training Quantization (PTQ)

python opencood/tools/inference_mc_quant.py ${CHECKPOINT_FOLDER} [--fusion_method intermediate] --num_cali_batches 16 --n_bits_w 8 --n_bits_a 8 --iters_w 5000
  • num_cali_batches refers to the size of the calibration dataset.
  • n_bits_w refers to the bitwidth for weight quantization.
  • n_bits_a refers to the bitwidth for activation quantization.
  • iters_w refers to the number of calibration steps.

Notes:

  • You could refer to /scripts/inference_mc/inference_mc_quant.sh for example running scripts. mc stands for multi-class, which differentiates itself from single-class training and inference. quant stands for the PTQ process that is different from fp which stands for full-precision inference.