分支列表

默认分支

71d316d608 · Deprecate unused DeepSpeed features (#8490) · 最后更新于 2026-09-11 22:50:44 +00:00

分支列表

b372147e5b · Update torch, switch MII and DS unit test order · 最后更新于 2024-10-28 21:35:40 +00:00    frostbyte_neo

870
8

61029f9ece · Update to version 1.18 firmware · 最后更新于 2024-10-28 18:03:39 +00:00    frostbyte_neo

871
1

0a78c7b716 · add option to disable mixed precision · 最后更新于 2024-10-28 04:07:36 +00:00    frostbyte_neo

875
12

714575208c · Merge branch 'master' into tohtana/clean_all_param_coordinators · 最后更新于 2024-10-23 16:50:03 +00:00    frostbyte_neo

879
7

429a7f22eb · Merge branch 'master' into jomayeri/aio-mem-fix · 最后更新于 2024-10-17 23:39:59 +00:00    frostbyte_neo

885
9

31ec0c2663 · Merge branch 'tohtana/file_store_for_tests' of github.com:microsoft/DeepSpeed into tohtana/file_store_for_tests · 最后更新于 2024-10-17 20:49:07 +00:00    frostbyte_neo

886
5

c0a08bb33b · update torch version in workflows · 最后更新于 2024-10-16 20:36:22 +00:00    frostbyte_neo

887
1

a57a2e944b · Merge branch 'master' into tohtana/lock_hf_cache_update · 最后更新于 2024-10-15 20:16:04 +00:00    frostbyte_neo

889
3

3abd24d001 · aio lock in param swapper · 最后更新于 2024-10-15 20:11:56 +00:00    frostbyte_neo

895
1

7b108a3f94 · Merge branch 'master' into tohtana/debug_semaphore_leak · 最后更新于 2024-10-15 16:54:38 +00:00    frostbyte_neo

890
24

137541f242 · cosolidating function · 最后更新于 2024-10-14 17:58:56 +00:00    frostbyte_neo

895
9

a41863bb60 · revert workflow · 最后更新于 2024-10-12 05:05:17 +00:00    frostbyte_neo

896
4

5c4b97f109 · apply fp16 autocast only to floating point values · 最后更新于 2024-10-11 19:41:10 +00:00    frostbyte_neo

896
0
已包含

dcf7af9a50 · remove unneeded comments and add zero 2 · 最后更新于 2024-10-10 07:41:51 +00:00    frostbyte_neo

911
17

c335571a8a · Merge branch 'master' into tohtana/get_offload_state_api · 最后更新于 2024-10-10 01:18:07 +00:00    frostbyte_neo

899
36

0182208339 · Merge branch 'master' into jomayeri/aio-locked-tensor · 最后更新于 2024-10-09 18:48:01 +00:00    frostbyte_neo

901
26
loadams/a6000-fix-0-15-2
Build and publish DeepSpeed release / deploy (push) Has been cancelled

474a3288cd · Enabled Qwen2-MoE Tensor Parallelism (TP) inference (#6551) · 最后更新于 2024-10-09 15:23:16 +00:00    frostbyte_neo

902
0
已包含

47af705560 · Merge branch 'master' into tohtana/clean_up_prefetch_param · 最后更新于 2024-10-09 00:20:11 +00:00    frostbyte_neo

906
37

5372ee6312 · Merge branch 'master' into olruwase/set_zero_opt_grad · 最后更新于 2024-10-08 18:39:02 +00:00    frostbyte_neo

909
8

430dcc2db6 · Update nv-lightning to run on Cuda 11.8/7? · 最后更新于 2024-10-07 23:32:44 +00:00    frostbyte_neo

912
1