fix

utils/interface.py

@@ -67,9 +67,7 @@ class Inference:
         # This should always be exactly one modality based on task
         if len(output_modalities) != 1: 
             raise ValueError("Expected exactly one output modality based on task.")
-        print(f"{[full_conversation]=}")
-        inputs = self.processor([full_conversation], output_modalities)
-
+        
         stopping_criteria = [
             MIMOStopper(self.processor.tokenizer.pad_token_id),
             MIMOStopper(
@@ -91,7 +89,8 @@ class Inference:
 
         @spaces.GPU(duration = 120)
         def gen_spaces():
-            print(f"{inputs['input_ids'].tolist()=}")
+            inputs = self.processor([full_conversation], output_modalities)
+
             token_ids = self.model.generate(
                 input_ids=inputs["input_ids"].to(self.device),
                 attention_mask=inputs["attention_mask"].to(self.device),