pytorch-image-models/models/__init__.py

from models.model_factory import create_model
from models.helpers import load_checkpoint, resume_checkpoint
from models.test_time_pool import TestTimePoolHead, apply_test_time_pool
Uniform pretrained model handling. * All models have 'default_cfgs' dict * load/resume/pretrained helpers factored out * pretrained load operates on state_dict based on default_cfg * test all models in validate * schedule, optim factor factored out * test time pool wrapper applied based on default_cfg 2019-04-12 12:32:16 +08:00			`from models.model_factory import create_model`
			`from models.helpers import load_checkpoint, resume_checkpoint`
			`from models.test_time_pool import TestTimePoolHead, apply_test_time_pool`
Add some Nvidia performance enhancements (prefetch loader, fast collate), and refactor some of training and model fact/transforms 2019-03-11 05:23:16 +08:00