Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Added sample cpu_offloading switch to YAML #8148

Merged
merged 5 commits into from
Jan 24, 2024
Merged

Added sample cpu_offloading switch to YAML #8148

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
c37d2a7
Added sample cpu_offloading switch to YAML
Jan 10, 2024
8ddfeeb
Added comments
Jan 10, 2024
3065ddf
Removed arithmetic op
Jan 23, 2024
380b70e
Merge branch 'main' into cpu_offload
ericharper Jan 23, 2024
2e183f6
Merge branch 'main' into cpu_offload
ericharper Jan 24, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions examples/nlp/language_modeling/conf/megatron_gpt_config.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -202,6 +202,12 @@ model:
## Flash Attention
use_flash_attention: False # Use flash attention in self-attention module, this config does nothing when transformer_engine=True

##Offloading Activations/Weights to CPU
cpu_offloading: False
cpu_offloading_num_layers: 11 #This value should be between [1,num_layers-1] as we don't want to offload the final layer's activations and expose any offloading duration for the final layer
cpu_offloading_activations: True
cpu_offloading_weights: True

## Network
sharp: False # Enable the use of SHARP for NCCL data-parallel communications. This is going to be ignored if the network doesn't support SHARP.

Expand Down
Loading