Pre-downscale factor

gradio_demo.py

@@ -125,6 +125,7 @@ def stage2_process(
     a_prompt,
     n_prompt,
     num_samples,
+    downscale,
     upscale,
     edm_steps,
     s_stage1,
@@ -152,6 +153,9 @@ def stage2_process(
         gr.Warning('Set this space to GPU config to make it work.')
         return None, None, None
     input_image = noisy_image if denoise_image is None else denoise_image
+    if 1 < downscale:
+        input_height, input_width, input_channel = np.array(input_image).shape
+        input_image = input_image.resize((input_width // downscale, input_height // downscale), Image.LANCZOS)
     torch.cuda.set_device(SUPIR_device)
     event_id = str(time.time_ns())
     event_dict = {'event_id': event_id, 'localtime': time.ctime(), 'prompt': prompt, 'a_prompt': a_prompt,
@@ -320,18 +324,18 @@ with gr.Blocks(title="SUPIR") as interface:
     gr.HTML(title_html)
 
     input_image = gr.Image(label="Input", show_label=True, type="numpy", height=600, elem_id="image-input")
-    prompt = gr.Textbox(label="Image description for LlaVa", value="", placeholder="A person, walking, in a town, Summer, photorealistic", lines=3, visible=False)
-    upscale = gr.Radio([1, 2, 3, 4, 5, 6, 7, 8], label="Upscale factor", info="Resolution x1 to x8", value=2, interactive=True)
-    a_prompt = gr.Textbox(label="Image description (optional)",
-                          info="Help the AI understand what the image represents; describe as much as possible",
-                          value='Cinematic, High Contrast, highly detailed, taken using a Canon EOS R '
-                                'camera, hyper detailed photo - realistic maximum detail, 32k, Color '
-                                'Grading, ultra HD, extreme meticulous detailing, skin pore detailing, '
-                                'hyper sharpness, perfect without deformations.',
-                          lines=3)
     with gr.Group():
+        prompt = gr.Textbox(label="Image description for LlaVa", value="", placeholder="A person, walking, in a town, Summer, photorealistic", lines=3, visible=False)
+        upscale = gr.Radio([1, 2, 3, 4, 5, 6, 7, 8], label="Upscale factor", info="Resolution x1 to x8", value=2, interactive=True)
+        a_prompt = gr.Textbox(label="Image description",
+                              info="Help the AI understand what the image represents; describe as much as possible",
+                              value='Cinematic, High Contrast, highly detailed, taken using a Canon EOS R '
+                                    'camera, hyper detailed photo - realistic maximum detail, 32k, Color '
+                                    'Grading, ultra HD, extreme meticulous detailing, skin pore detailing, '
+                                    'hyper sharpness, perfect without deformations.',
+                              lines=3)
         a_prompt_hint = gr.HTML("You can use a <a href='"'https://huggingface.co/spaces/MaziyarPanahi/llava-llama-3-8b'"'>LlaVa space</a> to auto-generate the description of your image.")
-    output_format = gr.Radio(["png", "webp", "jpeg", "gif", "bmp"], label="Image format for result", info="File extention", value="png", interactive=True)
+        output_format = gr.Radio(["png", "webp", "jpeg", "gif", "bmp"], label="Image format for result", info="File extention", value="png", interactive=True)
 
     with gr.Accordion("Pre-denoising (optional)", open=False):
         gamma_correction = gr.Slider(label="Gamma Correction", minimum=0.1, maximum=2.0, value=1.0, step=0.1)
@@ -355,6 +359,7 @@ with gr.Blocks(title="SUPIR") as interface:
         edm_steps = gr.Slider(label="Steps", info="lower=faster, higher=more details", minimum=1, maximum=200, value=default_setting.edm_steps if torch.cuda.device_count() > 0 else 1, step=1)
         num_samples = gr.Slider(label="Num Samples", info="Number of generated results", minimum=1, maximum=4 if not args.use_image_slider else 1
                                 , value=1, step=1)
+        downscale = gr.Radio([1, 2, 3, 4, 5, 6, 7, 8], label="Pre-downscale factor", info="Reducing blurred image reduce the process time", value=1, interactive=True)
         with gr.Row():
             with gr.Column():
                 model_select = gr.Radio(["v0-Q", "v0-F"], label="Model Selection", info="Q=Quality, F=Fidelity", value="v0-Q",
@@ -444,6 +449,7 @@ with gr.Blocks(title="SUPIR") as interface:
         a_prompt,
         n_prompt,
         num_samples,
+        downscale,
         upscale,
         edm_steps,
         s_stage1,