9 Commits

Author	SHA1	Message	Date
ver217	ce5a7dcab0	[zero] Update sharded model v2 using sharded param v2 (#323)	2022-03-08 18:18:06 +08:00
jiaruifang	4d07bffd77	using pytest parametrize	2022-03-08 15:10:21 +08:00
Jiarui Fang	cec05b25c9	[zero] update zero context init with the updated test utils (#327)	2022-03-08 14:45:01 +08:00
Jiarui Fang	29521cba0a	[zero] yet an improved sharded param (#311)	2022-03-04 15:49:23 +08:00
Jiarui Fang	2f6295bf78	[zero] polish shard strategy (#310) ... * init shard param from shape tuple * add more unitest for shard param * add set_payload method for ShardedParam * [zero] add shareded tensor class * polish code * add shard stratgy * move shard and gather logic to shard strategy from shard tensor. * polish code	2022-03-04 15:35:07 +08:00
Jiarui Fang	88496b5b31	[zero] a shard strategy in granularity of tensor (#307)	2022-03-04 11:59:35 +08:00
Jiarui Fang	408cba655b	[zero] sharded tensor (#305) ... * init shard param from shape tuple * add more unitest for shard param * add set_payload method for ShardedParam * [zero] add shareded tensor class * polish code	2022-03-04 10:46:13 +08:00
Jiarui Fang	6c78946fdd	Polish sharded parameter (#297) ... * init shard param from shape tuple * add more unitest for shard param * add more unittests to shareded param	2022-03-03 12:42:57 +08:00
Jiarui Fang	3280869358	Feature/zero (#279) ... * add zero1 (#209) * add zero1 * add test zero1 * update zero stage 1 develop (#212) * Implement naive zero3 (#240) * naive zero3 works well * add zero3 param manager * add TODOs in comments * add gather full param ctx * fix sub module streams * add offload * fix bugs of hook and add unit tests * fix bugs of hook and add unit tests (#252) * add gather full param ctx * fix sub module streams * add offload * fix bugs of hook and add unit tests * polish code and add state dict hook * fix bug * update unit test * refactor reconstructed zero code * clip_grad support zero3 and add unit test * add unit test for Zero3ParameterManager * [WIP] initialize the shard param class * [WIP] Yet another sharded model implementation (#274) * [WIP] initialize the shard param class * [WIP] Yes another implementation of shardModel. Using a better hook method. * torch.concat -> torch.cat * fix test_zero_level_1.py::test_zero_level_1 unitest * remove deepspeed implementation and refactor for the reconstructed zero module * polish zero dp unittests Co-authored-by: ver217 <lhx0217@gmail.com> Co-authored-by: Frank Lee <somerlee.9@gmail.com>	2022-03-01 18:17:01 +08:00