Fixed expected all tensors to be on the same device

app.py

@@ -12,6 +12,8 @@ else:
 tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-small")
 model = T5ForConditionalGeneration.from_pretrained("roborovski/superprompt-v1", torch_dtype=torch.float16)
 
+model.to(device)
+
 
 def generate(
     prompt, history, max_new_tokens=512, repetition_penalty=1.2, temperature=0.5, top_p=1, top_k=1, seed=42