legacy_megatron-lm/deepspeed_ZERO1&2 FP16_Optimizer wrapper #9184
FP16_Optimizer Wrap
0b26d835
refine the code
d7231476
pengwa
added component:optimizer
pengwa
removed component:optimizer
pengwa
added component:ortmodule
allow model parallelism aggregation optional
4a205d6c
zero1 optimizer inefficient function override
ffed3300
ytaous
commented
on 2021-09-28
add deepspeed zero1 and zero2 - checkoverflow & clip norm
bcbc09f3
refine the code
17e12a54
minor fix
6a554eac
refine the code after validation
0955a107
Merge branch 'master' of https://github.com/microsoft/onnxruntime int…
c50cd289
fix modifier
8b416a7e
fix build
82708a90
fix a indent
7c84e010
pengwa
changed the title FP16_Optimizer wrapper legacy_megatron-lm/deepspeed_ZERO1&2 FP16_Optimizer wrapper 4 years ago
Merge branch 'master' of https://github.com/microsoft/onnxruntime int…
6a71c5cf
Merge branch 'master' of https://github.com/microsoft/onnxruntime int…
0d180751
resolve comments
bd5d4eb0
update the document
2da5bc31
refine comment
7d1d4d17
refine with comments
5b5f4fae
pengwa
dismissed their stale review
via 5b5f4fae
4 years ago
refine after validation
71813f58
pengwa
merged
5ee47e3f
into master 4 years ago
pengwa
deleted the pengwa/opt_fp16_optimizer branch 4 years ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub