Upload 3 files

convert_ner_court_decisions.py

@@ -16,7 +16,7 @@ def proc_dataset(dataset, max_text_length=512):
                 before_tokens = random.randint(0, int(allowed_tokens * 0.8))
                 after_tokens = allowed_tokens - before_tokens
                 target_text = (" ".join(text.split(" ")[:20]) + " ... "
-                               + text[entity["start"] - before_tokens:entity["end"] + after_tokens])
+                               + " ".join(text.split(" ")[entity["start"] - before_tokens:entity["end"] + after_tokens]))
             r.append({
                 "label": entity["category_str"],
                 "answers": [entity["content"]],
@@ -28,8 +28,8 @@ def proc_dataset(dataset, max_text_length=512):
 d = load_dataset("fewshot-goes-multilingual/cs_czech-court-decisions-ner")
 train = list(d['train'])
 random.shuffle(train)
-new_dataset_train = proc_dataset(train[75:])
-dataset_test_ftrain = proc_dataset(train[:75])
+new_dataset_train = proc_dataset(train[60:])
+dataset_test_ftrain = proc_dataset(train[:60])
 dataset_val = proc_dataset(d['validation'])
 dataset_test = proc_dataset(d['test'])
 
@@ -41,6 +41,10 @@ random.shuffle(new_dataset_test)
 os.makedirs(".data/hf_dataset/ner_court_decisions", exist_ok=True)
 import jsonlines
 
+# print dataset lengths
+print("train", len(new_dataset_train))
+print("test", len(new_dataset_test))
+
 with jsonlines.open(".data/hf_dataset/ner_court_decisions/test.jsonl", "w") as f:
     f.write_all(new_dataset_test)
 with jsonlines.open(".data/hf_dataset/ner_court_decisions/train.jsonl", "w") as f: