Update handler.py

handler.py

@@ -28,11 +28,11 @@ class EndpointHandler():
         prompts = [f"<human>: {prompt}\n<bot>:" for prompt in data["inputs"]]
         
         
-        inputs = tokenizer(prompts, padding=True, return_tensors='pt').to(model.device)
+        inputs = self.tokenizer(prompts, padding=True, return_tensors='pt').to(model.device)
         input_length = inputs.input_ids.shape[1]
-        outputs = model.generate(
+        outputs = self.model.generate(
             **inputs, max_new_tokens=512, do_sample=True, temperature=0.7, top_p=0.7, top_k=50
          )
-        output_strs = tokenizer.batch_decode(outputs, skip_special_tokens=True)
+        output_strs = self.tokenizer.batch_decode(outputs, skip_special_tokens=True)
         
         return output_strs