Update app.py

app.py

@@ -23,7 +23,7 @@ def read_content(file_path: str) -> str:
 
     return content
 
-def predict(image, prompt="high quality, best quality", negative_prompt="", guidance_scale=5, steps=30, ip_adapter_scale = 1.0, width=1024, height=1024, seed=0):
+def predict(image, prompt="high quality, best quality", negative_prompt="", guidance_scale=5, steps=30, ip_adapter_scale = 1.0, width=1024, height=1024, seed=0, center_crop=False):
 
     pipe.set_ip_adapter_scale(ip_adapter_scale)
 
@@ -31,7 +31,8 @@ def predict(image, prompt="high quality, best quality", negative_prompt="", guid
         negative_prompt = None
     
     init_image = image.convert("RGB")
-    init_image = init_image.resize((224, 224))
+    if center_crop is False:
+        init_image = init_image.resize((224, 224))
     
     generator = torch.Generator(device="cpu").manual_seed(int(seed))
         
@@ -120,6 +121,7 @@ with image_blocks as demo:
                             steps = gr.Number(value=30, minimum=10, maximum=100, step=1, label="steps")
                             seed = gr.Number(value=0, minimum=0, maximum=100000, step=1, label="seed")                                                        
                             negative_prompt = gr.Textbox(label="negative_prompt", value=default_negative_prompt, placeholder=default_negative_prompt, info="what you don't want to see in the image")
+                            center_crop = gr.Checkbox(label="Center Crop", info="If not checked, the image would be resized to square before it's fed to the model."),
 
                         
                 with gr.Column():
@@ -127,8 +129,8 @@ with image_blocks as demo:
 
             
 
-    btn.click(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed], outputs=[image_out], api_name='run')
-    prompt.submit(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed], outputs=[image_out])
+    btn.click(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed], center_crop, outputs=[image_out], api_name='run')
+    prompt.submit(fn=predict, inputs=[image, prompt, negative_prompt, guidance_scale, steps, ip_adapter_scale, width, height, seed, center_crop], outputs=[image_out])
 
     # gr.Examples(
     #             examples=[