neural_compressor.strategy.auto
The auto tuning strategy.
Module Contents
Classes
The auto tuning strategy. |
- class neural_compressor.strategy.auto.AutoTuneStrategy(model, conf, q_dataloader=None, q_func=None, eval_func=None, eval_dataloader=None, eval_metric=None, resume=None, q_hooks=None)[source]
The auto tuning strategy.
There are three stages executed by auto strategy sequentially, and the tuning process ends once the condition meets the exit policy.