add examples

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -15,4 +15,5 @@ debugs/
 models
 !*/models
 .ipynb_checkpoints
-checkpoints
+checkpoints
+gradio_cached_examples

app.py

@@ -39,6 +39,58 @@ savedir        = os.path.join(basedir, "samples/Gradio", datetime.now().strftime
 savedir_sample = os.path.join(savedir, "sample")
 os.makedirs(savedir, exist_ok=True)
 
+EXAMPLES = [ # prompt, first frame, width, height, center crop, seed
+    ["timelapse at the snow land with aurora in the sky.", "example/example_01.png"],
+    ["fireworks.", "example/example_02.png"],
+    ["clown fish swimming through the coral reef.", "example/example_03.png"],
+    ["melting ice cream dripping down the cone.", "example/example_04.png"],
+]
+
+EXAMPLES_HIDDEN = {
+    "timelapse at the snow land with aurora in the sky.": ["example/example_01.png", 256, 256, True, 21800],
+    "fireworks.": ["example/example_02.png", 256, 256, True, 21800],
+    "clown fish swimming through the coral reef.": ["example/example_03.png", 256, 256, True, 21800],
+    "melting ice cream dripping down the cone.": ["example/example_04.png", 256, 256, True, 21800]
+}
+
+def update_and_resize_image(input_image_path, height_slider, width_slider, center_crop):
+    if input_image_path.startswith("http://") or input_image_path.startswith("https://"):
+        pil_image = Image.open(requests.get(input_image_path, stream=True).raw).convert('RGB')
+    else:
+        pil_image = Image.open(input_image_path).convert('RGB')
+    original_width, original_height = pil_image.size
+    
+    if center_crop:
+        crop_aspect_ratio = width_slider / height_slider
+        aspect_ratio = original_width / original_height
+        if aspect_ratio > crop_aspect_ratio:
+            new_width = int(crop_aspect_ratio * original_height)
+            left = (original_width - new_width) / 2
+            top = 0
+            right = left + new_width
+            bottom = original_height
+            pil_image = pil_image.crop((left, top, right, bottom))
+        elif aspect_ratio < crop_aspect_ratio:
+            new_height = int(original_width / crop_aspect_ratio)
+            top = (original_height - new_height) / 2
+            left = 0
+            right = original_width
+            bottom = top + new_height
+            pil_image = pil_image.crop((left, top, right, bottom))
+    
+    pil_image = pil_image.resize((width_slider, height_slider))
+    return gr.Image(value=np.array(pil_image))
+
+
+def get_examples(prompt_textbox, input_image):
+    input_image_path = EXAMPLES_HIDDEN[prompt_textbox][0]
+    width_slider = EXAMPLES_HIDDEN[prompt_textbox][1]
+    height_slider = EXAMPLES_HIDDEN[prompt_textbox][2]
+    center_crop = EXAMPLES_HIDDEN[prompt_textbox][3]
+    seed_textbox = EXAMPLES_HIDDEN[prompt_textbox][4]
+    input_image = update_and_resize_image(input_image_path, height_slider, width_slider, center_crop)
+    return prompt_textbox, input_image, input_image_path, width_slider, height_slider, center_crop, seed_textbox
+
 # config models
 pipeline = ConditionalAnimationPipeline.from_pretrained("TIGER-Lab/ConsistI2V", torch_dtype=torch.float16)
 pipeline.to("cuda")
@@ -89,6 +141,14 @@ def animate(
     frame_init_noise_level,
     seed_textbox
 ):
+    width_slider = int(width_slider)
+    height_slider = int(height_slider)
+    frame_stride = int(frame_stride)
+    sample_step_slider = int(sample_step_slider)
+    txt_cfg_scale_slider = float(txt_cfg_scale_slider)
+    img_cfg_scale_slider = float(img_cfg_scale_slider)
+    frame_init_noise_level = int(frame_init_noise_level)
+
     if pipeline is None:
         raise gr.Error(f"Please select a pretrained pipeline path.")
     if input_image_path == "":
@@ -190,44 +250,6 @@ def animate(
         
     return gr.Video(value=save_sample_path)
 
-
-# @spaces.GPU
-# def run_pipeline(
-#         pipeline,
-#         prompt_textbox,
-#         negative_prompt_textbox,
-#         first_frame,
-#         sample_step_slider,
-#         width_slider,
-#         height_slider,
-#         txt_cfg_scale_slider,
-#         img_cfg_scale_slider,
-#         frame_stride,
-#         use_frameinit,
-#         frame_init_noise_level,
-        
-#     ):
-#     first_frame = first_frame.to("cuda")
-#     sample = pipeline(
-#         prompt_textbox,
-#         negative_prompt       = negative_prompt_textbox,
-#         first_frames          = first_frame,
-#         num_inference_steps   = sample_step_slider,
-#         guidance_scale_txt    = txt_cfg_scale_slider,
-#         guidance_scale_img    = img_cfg_scale_slider,
-#         width                 = width_slider,
-#         height                = height_slider,
-#         video_length          = 16,
-#         noise_sampling_method = "pyoco_mixed",
-#         noise_alpha           = 1.0,
-#         frame_stride          = frame_stride,
-#         use_frameinit         = use_frameinit,
-#         frameinit_noise_level = frame_init_noise_level,
-#         camera_motion         = None,
-#     ).videos
-#     return sample
-
-
 def ui():
     with gr.Blocks(css=css) as demo:
         gr.Markdown(
@@ -287,34 +309,16 @@ def ui():
                         input_image = gr.Image(label="Input Image", interactive=True)
                         input_image.upload(fn=update_textbox_and_save_image, inputs=[input_image, height_slider, width_slider, center_crop], outputs=[input_image_path, input_image])
                         result_video = gr.Video(label="Generated Animation", interactive=False, autoplay=True)
-
-            def update_and_resize_image(input_image_path, height_slider, width_slider, center_crop):
-                if input_image_path.startswith("http://") or input_image_path.startswith("https://"):
-                    pil_image = Image.open(requests.get(input_image_path, stream=True).raw).convert('RGB')
-                else:
-                    pil_image = Image.open(input_image_path).convert('RGB')
-                original_width, original_height = pil_image.size
-                
-                if center_crop:
-                    crop_aspect_ratio = width_slider / height_slider
-                    aspect_ratio = original_width / original_height
-                    if aspect_ratio > crop_aspect_ratio:
-                        new_width = int(crop_aspect_ratio * original_height)
-                        left = (original_width - new_width) / 2
-                        top = 0
-                        right = left + new_width
-                        bottom = original_height
-                        pil_image = pil_image.crop((left, top, right, bottom))
-                    elif aspect_ratio < crop_aspect_ratio:
-                        new_height = int(original_width / crop_aspect_ratio)
-                        top = (original_height - new_height) / 2
-                        left = 0
-                        right = original_width
-                        bottom = top + new_height
-                        pil_image = pil_image.crop((left, top, right, bottom))
                 
-                pil_image = pil_image.resize((width_slider, height_slider))
-                return gr.Image(value=np.array(pil_image))
+            with gr.Row():
+                batch_examples = gr.Examples(
+                    examples=EXAMPLES,
+                    fn=get_examples,
+                    cache_examples=True,
+                    examples_per_page=4,
+                    inputs=[prompt_textbox, input_image],
+                    outputs=[prompt_textbox, input_image, input_image_path, width_slider, height_slider, center_crop, seed_textbox],
+                )
             
             preview_button.click(fn=update_and_resize_image, inputs=[input_image_path, height_slider, width_slider, center_crop], outputs=[input_image])
             input_image_path.submit(fn=update_and_resize_image, inputs=[input_image_path, height_slider, width_slider, center_crop], outputs=[input_image])
@@ -339,6 +343,7 @@ def ui():
                 ],
                 outputs=[result_video]
             )
+        
             
     return demo