Max Ryabinin 62652e1717 Add Switch Transformers-like RemoteMixtureOfExperts (#228) пре 4 година
..
layers 62652e1717 Add Switch Transformers-like RemoteMixtureOfExperts (#228) пре 4 година
__init__.py 62652e1717 Add Switch Transformers-like RemoteMixtureOfExperts (#228) пре 4 година
checkpoints.py 3024d381c5 Support learning rate schedulers in ExpertBackend (#196) пре 4 година
connection_handler.py 916c3db52d Move compression-related code to hivemind.utils.compression (#213) пре 4 година
dht_handler.py f132294edb Extract expert-specific methods from DHT (#192) пре 4 година
expert_backend.py ca5c7610ae Add tool for custom user experts (#189) пре 4 година
expert_uid.py 3024d381c5 Support learning rate schedulers in ExpertBackend (#196) пре 4 година
runtime.py 6128cbbd51 Add gradient clipping support to ExpertBackend (#214) пре 4 година
task_pool.py 6f8f192150 Improve Runtime exception handling (#207) пре 4 година