Merge pull request #397 from instructlab/courtneypacheco-fix-large-e2…

…e-job fix: `--keep_last_checkpoint_only` does not accept any values
instructlab · Jan 16, 2025 · 8969740 · 8969740
2 parents 989e4fb + 83bd68d
commit 8969740
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/src/instructlab/training/main_ds.py b/src/instructlab/training/main_ds.py
@@ -707,9 +707,11 @@ def run_training(torch_args: TorchrunArgs, train_args: TrainingArgs) -> None:
         f"--max_batch_len={train_args.max_batch_len}",
         f"--seed={train_args.random_seed}",
         f"--chat-tmpl-path={train_args.chat_tmpl_path}",
-        f"--keep_last_checkpoint_only={train_args.keep_last_checkpoint_only}",
     ]
 
+    if train_args.keep_last_checkpoint_only:
+        command.append(f"--keep_last_checkpoint_only")
+
     if train_args.checkpoint_at_epoch:
         command.append("--checkpoint_at_epoch")