From 7e3824f0cececdbee64f40a74b6681f99537839c Mon Sep 17 00:00:00 2001 From: Chao Pang Date: Mon, 16 Sep 2024 17:28:06 -0400 Subject: [PATCH] removed max_sequence_length from calculating the hash value for the prepared dataset --- src/cehrbert/runners/runner_util.py | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/src/cehrbert/runners/runner_util.py b/src/cehrbert/runners/runner_util.py index e6cf0f5e..4c94b699 100644 --- a/src/cehrbert/runners/runner_util.py +++ b/src/cehrbert/runners/runner_util.py @@ -187,9 +187,7 @@ def generate_prepared_ds_path(data_args, model_args, data_folder=None) -> Path: """ data_folder = data_folder if data_folder else data_args.data_folder concatenated_str = ( - str(model_args.max_position_embeddings) - + "|" - + os.path.abspath(data_folder) + +os.path.abspath(data_folder) + "|" + os.path.abspath(model_args.tokenizer_name_or_path) + "|"