18 Commits (5e5139283b03adf0b9dc092eb7b16a3076089769)

Author	SHA1	Message	Date
xjqbest	5e5139283b	fix runtime error ... test=develop	6 years ago
dongdaxiang	718ea6dbd5	fix fleet code style ... test=develop	6 years ago
xjqbest	a99c8d0c29	fix client to client communication bug ... test=develop	6 years ago
dongdaxiang	98dda08a85	fix pull sparse slow problem ... test=develop	6 years ago
dongdaxiang	d4514949bf	remove local random engine in fleet with rand_r() ... test=develop	6 years ago
dongdaxiang	365be5d559	support win32 flag in io.cc shell.cc, fix code style problem in fleet_wrapper, fix lodtensor_printer_test problem ... test=develop	6 years ago
xjqbest	b7940c2918	fix bug of gen_worker_desc and set_filelist, add some doc	6 years ago
xujiaqi01	f5c6a14b54	fix runtime error	6 years ago
xujiaqi01	a5b1a0e12b	support multi dataset && add init model && fix bug	6 years ago
dongdaxiang	b7a202aa38	add distributed optimizer factory	6 years ago
xujiaqi01	d25389fefd	add some log && fix error	6 years ago
dongdaxiang	317eb0aad3	add incubate for unified API	6 years ago
xujiaqi01	39449ba0b9	fix bug && add DestroyReaders in trainer	6 years ago
xujiaqi01	ecfc7df913	add dataset factory && fix style	6 years ago
xujiaqi01	3cea00bd52	store memory data in Dataset && fix bug	6 years ago
dongdaxiang	6de9ebc65c	refine VLOG in fleet_wrapper.h ... test=develop	6 years ago
dongdaxiang	c165012031	refine device_worker and trainer code ... test=develop	6 years ago
dongdaxiang	855bf579d2	add dist_multi_trainer for distributed training, add trainer_factory and device_worker_factory so that we can easily extend new training mode, add pull dense worker which is a singleton for parameter fetching	6 years ago