Skip to content

Train: add gradient clipping before optimizer step #287

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

Open
renee-jia wants to merge 1 commit into
base: master
Choose a base branch
from
+3 −0
Open

Train: add gradient clipping before optimizer step #287

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
3 changes: 3 additions & 0 deletions train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -445,6 +445,7 @@ def step(self):
WARMUP_RATIO = 0.0 # fraction of time budget for LR warmup
WARMDOWN_RATIO = 0.5 # fraction of time budget for LR warmdown
FINAL_LR_FRAC = 0.0 # final LR as fraction of initial
GRAD_CLIP_NORM = 1.0 # max gradient norm (0.0 to disable)

# Model size
DEPTH = 8 # number of transformer layers
Expand Down Expand Up @@ -561,6 +562,8 @@ def get_weight_decay(progress):
if group['kind'] == 'muon':
group["momentum"] = muon_momentum
group["weight_decay"] = muon_weight_decay
if GRAD_CLIP_NORM > 0.0:
torch.nn.utils.clip_grad_norm_(model.parameters(), GRAD_CLIP_NORM)
optimizer.step()
model.zero_grad(set_to_none=True)

Expand Down