quant_config_dynamic / quantization.yml
quantization's picture
Rename quantization.yaml to quantization.yml
1263f4d
raw
history blame contribute delete
235 Bytes
device: cpu
model:
framework: pytorch
name: bert
quantization:
approach: post_training_dynamic_quant
tuning:
accuracy_criterion:
relative: 0.2
exit_policy:
max_trials: 4
timeout: 0
random_seed: 9527
version: 1.0