Upload 2 files

handler.py

@@ -16,8 +16,10 @@ if IS_COMPILE:
 def compile_pipeline(pipe) -> Any:
     pipe.transformer.fuse_qkv_projections()
     pipe.transformer.to(memory_format=torch.channels_last)
-    #pipe.transformer = torch.compile(pipe.transformer, mode="reduce-overhead", fullgraph=False, dynamic=False, backend="inductor")
-    pipe.transformer = torch.compile(pipe.transformer, mode="max-autotune", fullgraph=False, dynamic=False, backend="inductor")
+    pipe.transformer = torch.compile(pipe.transformer, mode="reduce-overhead", fullgraph=False, dynamic=False, backend="inductor")
+    pipe.vae.fuse_qkv_projections()
+    pipe.vae.to(memory_format=torch.channels_last)
+    pipe.vae = torch.compile(pipe.vae, mode="reduce-overhead", fullgraph=False, dynamic=False, backend="inductor")
     return pipe
 
 class EndpointHandler:

requirements.txt

@@ -12,4 +12,3 @@ Pillow
 sentencepiece
 protobuf
 pytorch-lightning
-xformers