From b561bd844c9b7507ed1c6564f89380cd917e47a4 Mon Sep 17 00:00:00 2001
From: Jorge Martinez Gil <Jorge.Martinez-Gil@scch.at>
Date: Thu, 28 May 2026 17:34:29 +0200
Subject: [PATCH] fix: pass train_dataset and eval_dataset to
 CloneDetectionTrainer constructor

The Trainer.__init__ validates that eval_dataset is provided when
eval_strategy='epoch' is set. Datasets were previously only assigned
in run(), causing the ValueError before training could start.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
---
 scripts/run_clone_experiment.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/scripts/run_clone_experiment.py b/scripts/run_clone_experiment.py
index 4c7bb40..07363ee 100644
--- a/scripts/run_clone_experiment.py
+++ b/scripts/run_clone_experiment.py
@@ -114,6 +114,8 @@ def main() -> None:
     trainer = CloneDetectionTrainer(
         model=model,
         args=get_training_args(args.output_dir, **training_overrides),
+        train_dataset=train_ds,
+        eval_dataset=val_ds,
         data_collator=DataCollatorWithPadding(
             tokenizer=tokenizer,
             pad_to_multiple_of=8,