From d116bcaec98cb436d3b63d7e9bb150cb05afbb03 Mon Sep 17 00:00:00 2001 From: "l.gabrysiak" Date: Tue, 25 Feb 2025 12:20:29 +0100 Subject: [PATCH] poprawka c.d. --- hft.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/hft.py b/hft.py index 984f23a..dcfdf35 100644 --- a/hft.py +++ b/hft.py @@ -109,7 +109,8 @@ tokenizer = AutoTokenizer.from_pretrained(model_name) model = CustomModel.from_pretrained(model_name) # Przygotowanie datasetu -data = prepare_dataset("files", "file_catalog.json") +catalog_path = "file_catalog.json" +data = prepare_dataset("files", catalog_path) dataset = load_dataset("dict", data=data) tokenized_dataset = dataset.map(tokenize_function, batched=True, remove_columns=dataset["train"].column_names)