fix: use correct repo_root path for dataset files

This commit is contained in:
Christian Medina
2026-07-01 16:35:47 -04:00
parent 1095032801
commit 293f9caf65

View File

@@ -110,8 +110,8 @@ def train(config_path):
from datasets import load_dataset
import os
# Get dataset paths - assuming run from /home/workdir/artifacts
repo_root = Path(__file__).parent
# Get dataset paths - go up to project root
repo_root = Path(__file__).parent.parent.parent
train_path = str(repo_root / "training" / "data" / "train.jsonl")
test_path = str(repo_root / "training" / "data" / "test.jsonl")
print(f"Looking for dataset at: {train_path}")
@@ -206,7 +206,7 @@ def check_setup(config_path):
# Check dataset files
print("\n3. Dataset files:")
repo_root = Path(__file__).parent
repo_root = Path(__file__).parent.parent.parent # Go up to project root
train_path = repo_root / "training" / "data" / "train.jsonl"
test_path = repo_root / "training" / "data" / "test.jsonl"