fix parameter name in eval_mteb (#6)

- fix parameter name in eval_mteb (978fec22ab2d2f497d79c9554202a0d9ffd60618)


Co-authored-by: Bing <[email protected]>

scripts/eval_mteb.py

@@ -119,6 +119,7 @@ CMTEB_TASK_LIST = ['TNews', 'IFlyTek', 'MultilingualSentiment', 'JDReview', 'Onl
                    'T2Retrieval', 'MMarcoRetrieval', 'DuRetrieval', 'CovidRetrieval', 'CmedqaRetrieval', 'EcomRetrieval', 'MedicalRetrieval', 'VideoRetrieval',
                    'ATEC', 'BQ', 'LCQMC', 'PAWSX', 'STSB', 'AFQMC', 'QBQTC', 'STS22']
 
+
 MTEB_PL = [
     "CBD","PolEmo2.0-IN","PolEmo2.0-OUT","AllegroReviews","PAC","MassiveIntentClassification","MassiveScenarioClassification",
     "SICK-E-PL","PPC","CDSC-E","PSC","8TagsClustering","SICK-R-PL","CDSC-R","STS22",
@@ -405,6 +406,8 @@ class Wrapper:
         self._target_device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
         self.eod_id = self.tokenizer.convert_tokens_to_ids("<|endoftext|>")
         self.instruction = instruction
+        self.default_query = default_query
+        self.force_default = force_default
  
         if self.tokenizer.padding_side != 'right':
             logger.warning(f"Change tokenizer.padding_side from {self.tokenizer.padding_side} to right")
@@ -544,9 +547,9 @@ class Wrapper:
 
     def _tokenize(self, sentences: List[str], is_query: bool):
         
-        batch_dict = tokenizer(sentences, max_length=max_length - 1, return_attention_mask=False, padding=False, truncation=True)
-        batch_dict['input_ids'] = [input_ids + [tokenizer.eos_token_id] for input_ids in batch_dict['input_ids']]
-        batch_dict = tokenizer.pad(batch_dict, padding=True, return_attention_mask=True, return_tensors='pt')
+        batch_dict = self.tokenizer(sentences, max_length=self.max_seq_len - 1, return_attention_mask=False, padding=False, truncation=True)
+        batch_dict['input_ids'] = [input_ids + [self.tokenizer.eos_token_id] for input_ids in batch_dict['input_ids']]
+        batch_dict = self.tokenizer.pad(batch_dict, padding=True, return_attention_mask=True, return_tensors='pt')
         batch_dict['is_causal'] = False
         return batch_dict