Modelpublic

quantization/quant_config_dynamic

sourceHugging Faceupdated 4y agoView on Hugging Face

0likes

quantization.yml14 linesDownload Raw Back to root

1device: cpu2model:3  framework: pytorch4  name: bert5quantization:6  approach: post_training_dynamic_quant7tuning:8  accuracy_criterion:9    relative: 0.210  exit_policy:11    max_trials: 412    timeout: 013  random_seed: 952714version: 1.0