Update app.py

app.py

@@ -5,7 +5,7 @@ import torch
 import spaces
 
 # Load the model and processor
-model = gen_model.VLMForCausalLM.from_pretrained("unum-cloud/uform-gen")
+model = gen_model.VLMForCausalLM.from_pretrained("unum-cloud/uform-gen").to('cuda')
 processor = gen_model.VLMProcessor.from_pretrained("unum-cloud/uform-gen")
 
 @spaces.GPU
@@ -22,7 +22,7 @@ def generate_caption(image, prompt):
             max_new_tokens=128,
             eos_token_id=32001,
             pad_token_id=processor.tokenizer.pad_token_id
-        ).to('cuda')
+        )
 
     prompt_len = inputs["input_ids"].shape[1]
     decoded_text = processor.batch_decode(output[:, prompt_len:])[0]