Jeff Rasley 41db1c2f03 ZeRO-Offload release (#391) | 4 years ago | |
---|---|---|
.. | ||
BingBertSquad | 838f53b761 Switches BBS example to use mbsize=3 and gas=2 to fit in 16GB of memory. (#341) | 4 years ago |
Megatron_GPT2 | 41db1c2f03 ZeRO-Offload release (#391) | 4 years ago |
run_sanity_check.py | 53ac794702 reduce size of megatron tests (#223) | 4 years ago |