Update image generation pipeline in app.py to use new model and enable sequential CPU offload

app.py

@@ -24,9 +24,10 @@ from live_preview_helpers import flux_pipe_call_that_returns_an_iterable_of_imag
 llm_client = Client("Qwen/Qwen2.5-72B-Instruct")
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
-pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to(device)
+pipe = FluxPipeline.from_pretrained("sayakpaul/FLUX.1-merged", torch_dtype=torch.bfloat16).to(device)
 pipe.vae.enable_tiling()
 pipe.vae.enable_slicing()
+pipe.enable_sequential_cpu_offload() # offloads modules to CPU on a submodule level (rather than model level)
 torch.cuda.empty_cache()
 
 pipe.flux_pipe_call_that_returns_an_iterable_of_images = flux_pipe_call_that_returns_an_iterable_of_images.__get__(pipe)
@@ -87,7 +88,7 @@ def generate_item_image(object_t2i_prompt):
     for image in pipe.flux_pipe_call_that_returns_an_iterable_of_images(
         prompt=object_t2i_prompt,
         guidance_scale=3.5,
-        num_inference_steps=28,
+        num_inference_steps=8,
         width=1024,
         height=1024,
         generator=torch.Generator("cpu").manual_seed(0),