Skip to content

fix: directly save final ckpt in save_model_dir #626

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

Merged
dushyantbehl merged 1 commit into from
Nov 5, 2025
Merged

fix: directly save final ckpt in save_model_dir #626

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
23 changes: 15 additions & 8 deletions tuning/config/acceleration_configs/fast_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,16 +96,21 @@ def on_save(
Also saves the final model in save_model_dir if provided.
"""

def checkpoint(checkpoint_dir, save_dir):
hf_converted_output_dir = os.path.join(
save_dir, "hf_converted_checkpoint"
)
if os.path.exists(hf_converted_output_dir):
def checkpoint(checkpoint_dir, save_dir, is_intermediate: bool = True):
if is_intermediate:
hf_converted_output_dir = os.path.join(
save_dir, "hf_converted_checkpoint"
)
else:
hf_converted_output_dir = save_dir

if os.path.exists(hf_converted_output_dir) and is_intermediate:
# If the folder already exists
# we return, since this is possible to happen
# saving the checkpointing at the end of the training
return
os.mkdir(hf_converted_output_dir)

os.makedirs(hf_converted_output_dir, exist_ok=True)
try:
recover_safetensors_from_dcp(
checkpoint_dir,
Expand Down Expand Up @@ -165,8 +170,10 @@ def checkpoint(checkpoint_dir, save_dir):
and state.global_step == state.max_steps
):
if not os.path.exists(self.save_model_dir):
os.mkdir(self.save_model_dir)
checkpoint(checkpoint_dir, self.save_model_dir)
os.makedirs(self.save_model_dir, exist_ok=True)
checkpoint(
checkpoint_dir, self.save_model_dir, is_intermediate=False
)

callbacks.append(
ConvertAndSaveHFCheckpointAtEverySave(
Expand Down
Loading