Update handler.py

handler.py

@@ -32,10 +32,13 @@ class EndpointHandler:
 
         #for chat in prompts:
             #print(chat)
-        input_ids = self.tokenizer(prompt, return_tensors="pt").input_ids.to("cuda")
-        generated_ids = self.model.generate(input_ids, max_new_tokens=750, temperature=0.8, repetition_penalty=1.1, do_sample=True, eos_token_id=self.tokenizer.eos_token_id)
-        response = self.tokenizer.decode(generated_ids[0][input_ids.shape[-1]:], skip_special_tokens=True, clean_up_tokenization_space=True)
-        return response
+        
+        encodeds = self.tokenizer.encode(prompt, return_tensors="pt")
+        model_inputs = encodeds.to(device)
+        self.model.to(device)
+        generated_ids = self.model.generate(model_inputs, max_new_tokens=1000, do_sample=True)
+        decoded = self.tokenizer.decode(generated_ids[0])
+        return decoded
 
         """
         encodeds = self.tokenizer.encode(prompt, return_tensors="pt")