pytorch/ignite

Checkpoint stateful handlers and metrics

Open

#966 opened on 2020幎4月22日

GitHub で芋る
 (3 comments) (4 reactions) (0 assignees)Python (602 forks)batch import
enhancementhelp wantedmodule: metrics

Repository metrics

Stars
 (4,313 stars)
PR merge metrics
 (平均マヌゞ 18d 23h) (30d で 19 merged PRs)

説明

🚀 Feature

Things that are attached to the Engine might have state that would ideally be checkpointed and restored using as part of the Engine's state_dict.

An example is a Checkpoint handler when it has a score_function. Currently, the priorities the Checkpoint class stores is not saved anywhere. It is not able to recover gracefully from failure without manual intervention to parse the checkpoint path names, and directly setting the internals of the class.

Handlers and Metrics should have state_dict and load_state_dict methods (empty by default), and I think it should be possible for these to automatically make it into/restored from the Engine's state_dict when it's attached to an Engine.

コントリビュヌタヌガむド