changed chunk generation

ref_seg_ger.py

@@ -19,6 +19,7 @@ import os
 import numpy as np
 from PIL import Image
 from transformers import AutoTokenizer
+from tokenizers.pre_tokenizers import Whitespace
 import datasets
 from itertools import chain
 import pandas as pd
@@ -206,7 +207,7 @@ class RefSeg(datasets.GeneratorBasedBuilder):
         #print(filepath)
         #print(split)
         paths = glob(filepath + '/' + split + '/*.csv')
-        #print(paths)
+        print(paths)
         key = 0
         for f in paths:
             print(f)
@@ -215,7 +216,9 @@ class RefSeg(datasets.GeneratorBasedBuilder):
             labels = []
             for i, row in df.iterrows():
 
-                tokenized_input = row['token'].split(' ')
+                #tokenized_input = row['token'].split(' ')
+                tokenized_input, offsets = zip(*Whitespace(row['token']))
+                print(tokenized_input)
                 if f.endswith('Cermaine_0.xml.csv'):
                     print(tokenized_input)
                 if len(tokenized_input) > 1: