ceos/unsloth
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix: propagate is_embedding into worker subprocess config

Browse source 
start_training() cherry-picks kwargs into a config dict but was missing
is_embedding, so config.get("is_embedding", False) in worker.py always
returned False and embedding training never ran.
This commit is contained in:
Roland Tannous 2026-03-10 19:05:47 +00:00
commit d6ae910edc
1 changed files with 1 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

1
studio/backend/core/training/training.py
View file

@ -150,6 +150,7 @@ class TrainingBackend:
"custom_format_mapping": kwargs.get("custom_format_mapping"),
"is_dataset_image": kwargs.get("is_dataset_image", False),
"is_dataset_audio": kwargs.get("is_dataset_audio", False),
"is_embedding": kwargs.get("is_embedding", False),
"num_epochs": kwargs.get("num_epochs", 3),
"learning_rate": kwargs.get("learning_rate", "2e-4"),
"batch_size": kwargs.get("batch_size", 2),