fix: fix dataset path calculation
This commit is contained in:
@@ -61,10 +61,10 @@ def train(config_path):
|
|||||||
from datasets import load_dataset
|
from datasets import load_dataset
|
||||||
import os
|
import os
|
||||||
|
|
||||||
# Get dataset paths relative to script location
|
# Get dataset paths - dataset is in training/data/ relative to repo root
|
||||||
script_dir = Path(__file__).parent.parent
|
repo_root = Path(__file__).parent.parent.parent
|
||||||
train_path = str(script_dir / ".." / ".." / "training" / "data" / "train.jsonl")
|
train_path = str(repo_root / "training" / "data" / "train.jsonl")
|
||||||
test_path = str(script_dir / ".." / ".." / "training" / "data" / "test.jsonl")
|
test_path = str(repo_root / "training" / "data" / "test.jsonl")
|
||||||
|
|
||||||
dataset = load_dataset(
|
dataset = load_dataset(
|
||||||
"json",
|
"json",
|
||||||
|
|||||||
Reference in New Issue
Block a user