Update app.py

app.py

@@ -29,6 +29,11 @@ device = "cuda" if torch.cuda.is_available() else "cpu"
 
 torch.backends.cuda.matmul.allow_tf32 = True
 
+torch._inductor.config.conv_1x1_as_mm = True
+torch._inductor.config.coordinate_descent_tuning = True
+torch._inductor.config.epilogue_fusion = False
+torch._inductor.config.coordinate_descent_check_all_directions = True
+
 dtype = torch.bfloat16
 pipe = FluxWithCFGPipeline.from_pretrained("ostris/OpenFLUX.1", torch_dtype=dtype
 ).to("cuda")
@@ -54,6 +59,12 @@ pipe.tokenizer_max_length = maxtokens
 pipe.text_encoder.dtype = torch.bfloat16
 torch.cuda.empty_cache()
 
+pipe.transformer.to(memory_format=torch.channels_last)
+pipe.vae.to(memory_format=torch.channels_last)
+
+pipe.transformer = torch.compile(pipe.transformer, mode="max-autotune", fullgraph=True)
+pipe.vae.decode = torch.compile(pipe.vae.decode, mode="max-autotune", fullgraph=True)
+
 # Load LoRAs from JSON file
 with open('loras.json', 'r') as f:
     loras = json.load(f)