fix: auto-split dataset if split files missing

This commit is contained in:
Christian Medina
2026-06-30 19:54:50 -04:00
parent bbdbe91871
commit d5ba471b98

View File

@@ -60,8 +60,13 @@ if [ ! -f "dataset/combined_20k.jsonl" ]; then
exit 1 exit 1
fi fi
echo "Preparing dataset..." echo "Checking for train.jsonl and test.jsonl..."
if [ ! -f "training/data/train.jsonl" ] || [ ! -f "training/data/test.jsonl" ]; then
echo "Split files not found, creating them..."
python3 training/scripts/prepare_dataset.py --input dataset/combined_20k.jsonl --output training/data python3 training/scripts/prepare_dataset.py --input dataset/combined_20k.jsonl --output training/data
else
echo "Split files already exist, using them."
fi
echo "" echo ""
echo "=== Starting Training ===" echo "=== Starting Training ==="