Skip to content

Commit

Permalink
dataset config should be dict
Browse files Browse the repository at this point in the history
  • Loading branch information
milocress committed Apr 23, 2024
1 parent 9cdc7a4 commit 08814e1
Showing 1 changed file with 1 addition and 2 deletions.
3 changes: 1 addition & 2 deletions llmfoundry/data/text_data.py
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,6 @@
import transformers
from composer.core.data_spec import DataSpec
from composer.core.types import Batch
from omegaconf import DictConfig
from omegaconf import OmegaConf as om
from streaming import Stream, StreamingDataset
from torch.utils.data import DataLoader
Expand Down Expand Up @@ -267,7 +266,7 @@ def build_streams(streams: Optional[Dict[str, Any]] = None,):
def build_text_dataloader(
tokenizer: PreTrainedTokenizerBase,
device_batch_size: int,
dataset: DictConfig,
dataset: Dict[str, Any],
drop_last: bool,
num_workers: int,
pin_memory: bool = True,
Expand Down

0 comments on commit 08814e1

Please sign in to comment.