use pie-modules instead of pytorch-ie

see https://github.com/ArneBinder/pie-datasets/pull/204 for further information

chemprot.py

@@ -2,9 +2,9 @@ from dataclasses import dataclass
 from typing import Any, Dict
 
 import datasets
-from pytorch_ie import Document
-from pytorch_ie.annotations import BinaryRelation, LabeledSpan
-from pytorch_ie.documents import (
+from pie_core import Document
+from pie_modules.annotations import BinaryRelation, LabeledSpan
+from pie_modules.documents import (
     AnnotationLayer,
     TextBasedDocument,
     TextDocumentWithLabeledSpansAndBinaryRelations,

requirements.txt

@@ -1 +1,2 @@
-pie-datasets>=0.6.0,<0.11.0
+pie-datasets>=0.10.11,<0.11.0
+pie-modules>=0.15.9,<0.16.0