Konrad Wojtasik commited on
Commit
ade42bf
1 Parent(s): f1acfeb

Load all corpus

Browse files
Files changed (1) hide show
  1. app.py +1 -1
app.py CHANGED
@@ -27,7 +27,7 @@ auth_token = os.environ.get("auth_token")
27
  def load_data(dataset_type):
28
 
29
  corpus, queries, qrels = HFDataLoader(hf_repo="clarin-knext/"+dataset_type, streaming=False, keep_in_memory=False).load(split="test")
30
- corpus = [ doc['text']for doc in corpus][:100]
31
  queries = [ query['text']for query in queries]
32
  return queries, corpus
33
 
 
27
  def load_data(dataset_type):
28
 
29
  corpus, queries, qrels = HFDataLoader(hf_repo="clarin-knext/"+dataset_type, streaming=False, keep_in_memory=False).load(split="test")
30
+ corpus = [ doc['text']for doc in corpus]
31
  queries = [ query['text']for query in queries]
32
  return queries, corpus
33