roytseng-tw / Detectron.pytorch

A pytorch implementation of Detectron. Both training from scratch and inferring directly from pretrained Detectron weights are available.
MIT License
2.82k stars 565 forks source link

Fix resume error #118

Closed ncherel closed 6 years ago

ncherel commented 6 years ago

Fixes #109 named_parameters is an ordered dict and the order can be lost when converting to a dict, resulting in shape mismatches between different runs. The default state_dict loader can then be used.

Warning : only tested with PyTorch 0.4.0