Datasets:

Formats:
parquet
Languages:
English
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
License:
robiks commited on
Commit
0cb4bfe
·
verified ·
1 Parent(s): 87abe93

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -1
README.md CHANGED
@@ -157,7 +157,11 @@ dataset_info:
157
  dataset_size: 1276040
158
  ---
159
 
160
- The RealKIE-FCC-Verified test set includes single and multi-page invoices sourced from the Federal Communications Commission (FCC) to evaluate key information extraction (KIE) performance. It contains 75 documents, with: a) a common JSON schema and b) corresponding expected JSON responses. This release resolves two issues present in the original evaluation set (https://arxiv.org/abs/2403.20101):
 
 
 
 
161
 
162
  a) Line Item Grouping: Fields previously treated as independent entries are now grouped within each individual line item, aligning the annotations with real-world invoice structure.
163
 
 
157
  dataset_size: 1276040
158
  ---
159
 
160
+ The RealKIE-FCC-Verified test set includes single and multi-page invoices sourced from the Federal Communications Commission (FCC) to evaluate key information extraction (KIE) performance.
161
+
162
+ The task is to extract information from the document in JSON format given the corresponding JSON schema. It contains 75 documents, with: a) a common JSON schema and b) corresponding expected JSON responses.
163
+
164
+ This release resolves two issues present in the original evaluation set (https://arxiv.org/abs/2403.20101):
165
 
166
  a) Line Item Grouping: Fields previously treated as independent entries are now grouped within each individual line item, aligning the annotations with real-world invoice structure.
167