bug + only new tokens

handler.py

@@ -79,11 +79,12 @@ class EndpointHandler():
 
         # Generate text using the pipeline
         generation_kwargs = {
-            "max_length": 75,  # Adjust as needed
+            "max_new_tokens": 75,
             "temperature": 0.7,
             "top_k": 40,
             "bad_words_ids": bad_words_ids,
-            # "pad_token_id": self.tokenizer.eos_token_id  # Ensure padding with EOS token
+            "pad_token_id": self.tokenizer.eos_token_id,
+            "return_full_text": False,  # Only return the new generated tokens
         }
         generated_outputs = self.text_generation_pipeline(inputs, **generation_kwargs)