No CUDA support in hf

app.py

@@ -46,11 +46,11 @@ def respond(message, history: list[tuple[str, str]], system_message, max_tokens,
         tokenize = True,
         add_generation_prompt = True, # Must add for generation
         return_tensors = "pt",
-    ).to("cuda")
+    )
 
     from transformers import TextStreamer
     text_streamer = TextStreamer(tokenizer, skip_prompt = True)
-    
+
     yield model.generate(input_ids = inputs, streamer = text_streamer, max_new_tokens = 128,
                     use_cache = True, temperature = 1.5, min_p = 0.1)