Skip to content

"[QEff. Finetune]: Support for resuming checkpoints using Epoch" #614

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
merged 8 commits into from
Nov 11, 2025
Merged

"[QEff. Finetune]: Support for resuming checkpoints using Epoch" #614

Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
977f278
Adding support to load checkpoints from epoch
tchawada Nov 4, 2025
9ce2d13
Adding support to load checkpoints from epoch
tchawada Nov 4, 2025
4fb76eb
Adding support to load checkpoints from epoch
tchawada Nov 5, 2025
5443ac6
Adding support to load checkpoints from epoch
tchawada Nov 5, 2025
e6c3004
Merge branch 'main' into finetuning_checkpoint
tchawada Nov 5, 2025
c9f15f6
Merge branch 'quic:main' into finetuning_checkpoint
tchawada Nov 10, 2025
eea687c
[QEff. Finetune]:Adding support to load checkpoints from epoch
tchawada Nov 10, 2025
3d7ce0a
[QEff. Finetune]:Adding support to load checkpoints from epoch
tchawada Nov 10, 2025
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
9 changes: 5 additions & 4 deletions QEfficient/finetune/utils/train_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,11 +123,12 @@ def train(
break

if train_config.use_peft and train_config.from_peft_checkpoint:
path = train_config.from_peft_checkpoint.rstrip("/")
try:
intermediate_epoch = int(train_config.from_peft_checkpoint.split("/")[-2].split("_")[-1]) - 1
intermediate_step = int(train_config.from_peft_checkpoint.split("/")[-1].split("_")[-1])
intermediate_epoch = int(path.split("/")[-2].split("_")[-1]) - 1
intermediate_step = int(path.split("/")[-1].split("_")[-1])
except (IndexError, ValueError):
intermediate_epoch = int(train_config.from_peft_checkpoint.split("/")[-1].split("_")[-1]) - 1
intermediate_epoch = int(path.split("/")[-1].split("_")[-1]) - 1
intermediate_step = 0

if epoch < intermediate_epoch:
Expand Down Expand Up @@ -374,7 +375,7 @@ def train(
eval_step_metric,
eval_metric,
)
avg_epoch_time = sum(epoch_times) / len(epoch_times)
avg_epoch_time = sum(epoch_times) / len(epoch_times) if len(epoch_times) > 0 else 0
avg_checkpoint_time = sum(checkpoint_times) / len(checkpoint_times) if len(checkpoint_times) > 0 else 0

results["last_epoch_train_loss"] = train_epoch_loss.cpu()
Expand Down
Loading