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

add use_fp32_logits flag #393

Open
wants to merge 1 commit into
base: develop
Choose a base branch
from
Open

add use_fp32_logits flag #393

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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 internlm/model/ops/cross_entropy.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,12 @@ def new_cross_entropy(
except KeyError:
raise KeyError(f"op_type only support: {cross_entropy_op_name_map.keys()}")

if not gpc.config.get("use_fp32_logits", True):
assert op_type in [
CrossEntropyOpType.flash_vocab_parallel,
CrossEntropyOpType.apex_naive,
], "use_fp32_logits=False only support 'flash_vocab_parallel' or 'apex_naive' loss function"

if internlm_accelerator.get_accelerator_backend() is not AcceleratorType.GPU:
assert op_type in [
CrossEntropyOpType.torch_naive,
Expand Down
2 changes: 1 addition & 1 deletion internlm/train/pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -300,7 +300,7 @@ def inject_model(model):
else:
model = NaiveAMPModel(
model=model,
output_to_fp32=gpc.is_no_pp_or_last_stage(),
output_to_fp32=gpc.is_no_pp_or_last_stage() and gpc.config.get("use_fp32_logits", True),
dtype=gpc.config.model.get("dtype", torch.half),
sync_buffer=False,
)
Expand Down
Loading