Update app.py

app.py

@@ -18,7 +18,7 @@ repo_name = "alimama-creative/FLUX.1-Turbo-Alpha"
 ckpt_name = "diffusion_pytorch_model.safetensors"
 turbo_lora = hf_hub_download(repo_name, ckpt_name)
 
-pipe_dev = DiffusionPipeline.from_pretrained(dev_model, torch_dtype=torch.bfloat16)
+pipe_dev = DiffusionPipeline.from_pretrained(dev_model, torch_dtype=torch.bfloat16).to("cuda")
 pipe_schnell = DiffusionPipeline.from_pretrained(
     schnell_model,
     text_encoder=pipe_dev.text_encoder,
@@ -63,7 +63,6 @@ def run_schnell(prompt):
 
 @spaces.GPU
 def run_parallel_models(prompt):
-    pipe_dev.to("cuda")
     pipe_dev.load_lora_weights(hyper_lora)
     image = pipe_dev(prompt, num_inference_steps=8, joint_attention_kwargs={"scale": 0.125}).images[0]
     pipe_dev.unload_lora_weights()