ghndrx/kubeflow-pipelines
1
0
Fork 0
mirror of https://github.com/ghndrx/kubeflow-pipelines.git synced 2026-02-10 06:45:13 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: disable checkpoint saving to avoid tensor contiguity error

Browse Source
This commit is contained in:
Greg Hendrickson
2026-02-03 02:38:15 +00:00
parent 2680ad5502
commit 0f4858d22f
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
components/runpod_trainer/handler.py
View File

@@ -103,7 +103,7 @@ def train_ddi_model(job_input: Dict[str, Any]) -> Dict[str, Any]:
warmup_steps=50, warmup_steps=50,
weight_decay=0.01, weight_decay=0.01,
logging_steps=10, logging_steps=10,
save_strategy='epoch', save_strategy='no', # Don't save checkpoints (avoids tensor contiguity issues)
fp16=torch.cuda.is_available(), fp16=torch.cuda.is_available(),
report_to='none', report_to='none',
) )