Upload 82 files

.gitattributes

@@ -33,3 +33,34 @@ 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
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/1/frame_0001_2T3_real_esrgan_anime_x2.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/1/frame_0001_2T3.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/2/frame_0001_9T3_real_esrgan_anime_x4.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/2/frame_0001_9T3.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/3/frame_0001_10T3_real_esrgan_anime_x6.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/3/frame_0001_10T3.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/4/frame_0001_14T3_real_esrgan_anime_x8.png filter=lfs diff=lfs merge=lfs -text
+apps/assets/examples/Real-ESRGAN-Anime-finetuning/4/frame_0001_14T3.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/1/1_x2_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/1/1_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/1/frame_0001_2T3_real_esrgan_anime_x2.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/1/frame_0001_2T3.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/2/2_x4_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/2/2_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/2/frame_0001_9T3_real_esrgan_anime_x4.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/2/frame_0001_9T3.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/3/3_x6_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/3/3_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/3/frame_0001_10T3_real_esrgan_anime_x6.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/3/frame_0001_10T3.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/4/4_x8_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/4/4_zoom.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/4/frame_0001_14T3_real_esrgan_anime_x8.png filter=lfs diff=lfs merge=lfs -text
+assets/examples/Real-ESRGAN-Anime-finetuning/4/frame_0001_14T3.png filter=lfs diff=lfs merge=lfs -text
+tests/test_data/frame_0001_real_esrgan_anime.jpg filter=lfs diff=lfs merge=lfs -text
+tests/test_data/frame_0001_real_esrgan.jpg filter=lfs diff=lfs merge=lfs -text
+tests/test_data/image_x4.png filter=lfs diff=lfs merge=lfs -text
+tests/test_data/image_x8.png filter=lfs diff=lfs merge=lfs -text
+tests/test_data/input_image.png filter=lfs diff=lfs merge=lfs -text
+tests/test_data/real_esrgan_anime/frame_0235_chunk0_8T1_out.png filter=lfs diff=lfs merge=lfs -text
+tests/test_data/real_esrgan/frame_0235_chunk0_8T1_out.png filter=lfs diff=lfs merge=lfs -text

.python-version

@@ -0,0 +1 @@
+3.11.13

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Danh Tran
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

apps/assets/examples/Real-ESRGAN-Anime-finetuning/1/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_2T3.png",
+    "output_file": "frame_0001_2T3_real_esrgan_anime_x2.png",
+    "outer_scale": 2
+}

apps/assets/examples/Real-ESRGAN-Anime-finetuning/2/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_9T3.png",
+    "output_file": "frame_0001_9T3_real_esrgan_anime_x4.png",
+    "outer_scale": 4
+}

apps/assets/examples/Real-ESRGAN-Anime-finetuning/3/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_10T3.png",
+    "output_file": "frame_0001_10T3_real_esrgan_anime_x6.png",
+    "outer_scale": 6
+}

apps/assets/examples/Real-ESRGAN-Anime-finetuning/4/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_14T3.png",
+    "output_file": "frame_0001_14T3_real_esrgan_anime_x8.png",
+    "outer_scale": 8
+}

apps/gradio_app.py

@@ -0,0 +1,103 @@
+import gradio as gr
+import os
+import subprocess
+from gradio_app.project_info import (
+    CONTENT_DESCRIPTION,
+    CONTENT_IN_1, CONTENT_IN_2,
+    CONTENT_OUT_1, CONTENT_OUT_2
+)
+from gradio_app.inference import run_inference
+from gradio_app.examples import load_examples, select_example
+
+
+def run_setup_script():
+    setup_script = os.path.join(os.path.dirname(__file__),
+                                "gradio_app", "setup_scripts.py")
+    try:
+        result = subprocess.run(["python", setup_script], capture_output=True, text=True, check=True)
+        return result.stdout
+    except subprocess.CalledProcessError as e:
+        print(f"Setup script failed with error: {e.stderr}")
+        return f"Setup script failed: {e.stderr}"
+    
+def create_gui():
+    # Load custom CSS
+    custom_css = open("apps/gradio_app/static/styles.css").read()
+
+    # JavaScript function to update warning_text Markdown component
+    outer_scale_warning = open("apps/gradio_app/static/outer_scale_warning.js").read()
+
+    # Define Gradio interface
+    with gr.Blocks(css=custom_css) as demo:
+        gr.Markdown("# Anime Super Resolution 🖼️")
+        gr.Markdown(CONTENT_DESCRIPTION)
+        gr.Markdown(CONTENT_IN_1)
+        gr.HTML(CONTENT_IN_2)
+        with gr.Row():
+            with gr.Column(scale=2):
+                input_image = gr.Image(
+                    type="filepath",
+                    label="Input Image",
+                    elem_classes="input-image"
+                )
+                model_id = gr.Textbox(
+                    label="Model ID",
+                    value="danhtran2mind/Real-ESRGAN-Anime-finetuning"
+                )
+                
+                outer_scale = gr.Slider(
+                    minimum=1,
+                    maximum=8,
+                    step=1,
+                    value=2,
+                    label="Outer Scale",
+                    elem_id="outer-scale-slider"
+                )
+                warning_text = gr.HTML(elem_id="warning-text")
+                gr.Markdown(
+                    "**Note:** For optimal output quality, set `Outer Scale` to a value between 1 and 4. "
+                    "**Values greater than 4 are not recommended**. "
+                    "Please ensure `Outer Scale` is greater than or equal to `Inner Scale` (default: 4)."
+                )
+                
+                examples_data = load_examples()
+                submit_button = gr.Button("Run Inference")
+            
+            with gr.Column(scale=3):
+                output_image = gr.Image(
+                    label="Output Image",
+                    elem_classes="output-image"
+                )
+                output_text = gr.Textbox(label="Status")
+        
+        # Client-side warning update for warning_text
+        outer_scale.change(
+            fn=lambda x: x,
+            inputs=outer_scale,
+            outputs=outer_scale,
+            js=outer_scale_warning
+        )
+        
+        gr.Examples(
+            examples=[[input_img, output_img, outer_scale] for input_img, output_img, outer_scale in examples_data],
+            inputs=[input_image, output_image, outer_scale],
+            label="Example Inputs",
+            examples_per_page=4,
+            fn=select_example,
+            outputs=[input_image, outer_scale, output_image, output_text]
+        )
+            
+        submit_button.click(
+            fn=run_inference,
+            inputs=[input_image, model_id, outer_scale],
+            outputs=[output_image, output_text]
+        )
+        gr.HTML(CONTENT_OUT_1)
+        gr.Markdown(CONTENT_OUT_2)
+
+        return demo
+    
+if __name__ == "__main__":
+    run_setup_script()
+    demo = create_gui()
+    demo.launch()

apps/gradio_app/examples.py

@@ -0,0 +1,34 @@
+import os
+import json
+from PIL import Image
+import gradio as gr
+
+def load_examples():
+    examples = []
+    examples_base_path = os.path.join("apps", "assets", "examples", "Real-ESRGAN-Anime-finetuning")
+    
+    for folder in ["1", "2", "3", "4"]:
+        folder_path = os.path.join(examples_base_path, folder)
+        config_path = os.path.join(folder_path, "config.json")
+        
+        if os.path.exists(config_path):
+            with open(config_path, 'r') as f:
+                config = json.load(f)
+                input_filename = config.get("input_file", "input.jpg")
+                output_filename = config.get("output_file", "output.jpg")
+                outer_scale = config.get("outer_scale", 4)
+                
+                input_image_path = os.path.join(folder_path, input_filename)
+                output_image_path = os.path.join(folder_path, output_filename)
+                
+                if os.path.exists(input_image_path) and os.path.exists(output_image_path):
+                    input_image_data = Image.open(input_image_path)
+                    output_image_data = Image.open(output_image_path)
+                    examples.append([input_image_data, output_image_data, outer_scale])
+    
+    return examples
+
+def select_example(evt: gr.SelectData, examples_data):
+    example_index = evt.index
+    input_image_data, output_image_data, outer_scale = examples_data[example_index]
+    return input_image_data, outer_scale, output_image_data, f"Loaded example with Outer Scale: {outer_scale}"

apps/gradio_app/inference.py

@@ -0,0 +1,25 @@
+import gradio as gr
+from PIL import Image
+import os
+import sys
+
+# Add the project root directory to the Python path
+sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..')))
+
+from src.anime_super_resolution.infer import infer
+
+def run_inference(input_image, model_id, outer_scale):
+    if not input_image:
+        return None, "Please upload an image."
+    
+    models_config_path = "configs/model_ckpts.yaml"
+    try:
+        output_image = infer(
+            input_path=input_image,
+            model_id=model_id,
+            models_config_path=models_config_path,
+            outer_scale=outer_scale,
+        )
+        return output_image, "Inference completed successfully!"
+    except Exception as e:
+        return None, f"Error during inference: {str(e)}"

apps/gradio_app/project_info.py

@@ -0,0 +1,70 @@
+CONTENT_DESCRIPTION = """
+**Anime-Super-Resolution: High-Quality Anime Image Upscaling with Real-ESRGAN**
+""".strip()
+
+CONTENT_IN_1 = """
+Transforms low-res anime images into sharp, vibrant HD visuals, enhancing textures and details for artwork and games.
+""".strip()
+
+CONTENT_IN_2 = """
+<p class="source">
+    For more information, you can check out my GitHub repo:<br>
+    Source code: 
+    <a class="badge" href="https://github.com/danhtran2mind/Anime-Super-Resolution">
+        <img src="https://img.shields.io/badge/GitHub-danhtran2mind%2FAnime--Super--Resolution-blue?style=flat?logo=github" alt="GitHub Repo">
+    </a>,
+    Model Hub:
+    <a class="badge" href="https://huggingface.co/danhtran2mind/Real-ESRGAN-Anime-finetuning">
+        <img src="https://img.shields.io/badge/HuggingFace-danhtran2mind%2FReal--ESRGAN--Anime--finetuning-yellow?style=flat?logo=huggingface" alt="HuggingFace Model">
+    </a>.
+</p>
+""".strip()
+
+CONTENT_OUT_1 = """
+This project leverages code from:
+<div class="quote-container">
+    <p>
+        The Original Real-ESRGAN 
+        <a class="badge" href="https://github.com/xinntao/Real-ESRGAN">
+            <img src="https://img.shields.io/badge/Built%20on-xinntao%2FReal--ESRGAN-blue?style=flat&logo=github" alt="Built on Real-ESRGAN">
+        </a>
+        by 
+        <a class="badge" href="https://github.com/xinntao">
+            <img src="https://img.shields.io/badge/GitHub-xinntao-blue?style=flat&logo=github" alt="xinntao">
+        </a>, 
+        with our own bug fixes and enhancements available at 
+        <a class="badge" href="https://github.com/danhtran2mind/Real-ESRGAN">
+            <img src="https://img.shields.io/badge/GitHub-danhtran2mind%2FReal--ESRGAN-blue?style=flat&logo=github" alt="Real-ESRGAN Enhancements">
+        </a>.
+        <br>
+        The Inference code
+        <a class="badge" href="https://github.com/ai-forever/Real-ESRGAN">
+            <img src="https://img.shields.io/badge/Built%20on-ai--forever%2FReal--ESRGAN-blue?style=flat&logo=github" alt="Built on Real-ESRGAN">
+        </a>
+        by 
+        <a class="badge" href="https://github.com/ai-forever">
+            <img src="https://img.shields.io/badge/GitHub-ai--forever-blue?style=flat&logo=github" alt="xinntao">
+        </a>, 
+        with our own bug fixes and enhancements available at 
+        <a class="badge" href="https://github.com/danhtran2mind/Real-ESRGAN-inference">
+            <img src="https://img.shields.io/badge/GitHub-danhtran2mind%2FReal--ESRGAN--inference-blue?style=flat&logo=github" alt="Real-ESRGAN Enhancements">
+        </a>.
+    </p>
+</div>
+""".strip()
+
+CONTENT_OUT_2 = """
+You can explore more Model Hubs at:
+<div class="quote-container">
+    <p>
+        HuggingFace Model Hub: 
+        <a class="badge" href="https://huggingface.co/ai-forever/Real-ESRGAN">
+            <img src="https://img.shields.io/badge/HuggingFace-ai--forever%2FReal--ESRGAN-yellow?style=flat&logo=hugggingface" alt="ai-forever Real-ESRGAN Model">
+        </a>,
+        Real-ESRGAN Model releases: 
+        <a class="badge" href="https://github.com/xinntao/Real-ESRGAN/releases">
+            <img src="https://img.shields.io/badge/GitHub-Real--ESRGAN%2Freleases-blue?style=flat&logo=github" alt="Real-ESRGAN releases">
+        </a>.
+    </p>    
+</div>
+""".strip()

apps/gradio_app/setup_scripts.py

@@ -0,0 +1,63 @@
+import subprocess
+import sys
+import os
+
+def run_script(script_path, args=None):
+    """
+    Run a Python script using subprocess with optional arguments and handle errors.
+    Returns True if successful, False otherwise.
+    """
+    if not os.path.isfile(script_path):
+        print(f"Script not found: {script_path}")
+        return False
+
+    try:
+        command = [sys.executable, script_path]
+        if args:
+            command.extend(args)
+        result = subprocess.run(
+            command,
+            check=True,
+            text=True,
+            capture_output=True
+        )
+        print(f"Successfully executed {script_path}")
+        print(result.stdout)
+        return True
+    except subprocess.CalledProcessError as e:
+        print(f"Error executing {script_path}:")
+        print(e.stderr)
+        return False
+    except Exception as e:
+        print(f"Unexpected error executing {script_path}: {str(e)}")
+        return False
+
+def main():
+    """
+    Main function to execute download_ckpts.py with proper error handling.
+    """
+    scripts_dir = "scripts"
+    scripts = [
+        {
+            "path": os.path.join(scripts_dir, "download_ckpts.py"),
+            "args": []  # Empty list for args to avoid NoneType issues
+        },
+        {
+            "path": os.path.join(scripts_dir, "setup_third_party.py"),
+            "args": []
+        }
+    ]
+
+    for script in scripts:
+        script_path = script["path"]
+        args = script.get("args", [])  # Safely get args with default empty list
+        print(f"Starting execution of {script_path}{' with args: ' + ' '.join(args) if args else ''}\n")
+        
+        if not run_script(script_path, args):
+            print(f"Stopping execution due to error in {script_path}")
+            sys.exit(1)
+        
+        print(f"Completed execution of {script_path}\n")
+
+if __name__ == "__main__":
+    main()

apps/gradio_app/static/outer_scale_warning.js

@@ -0,0 +1,11 @@
+(value) => {
+    const warningElement = document.getElementById('warning-text');
+    if (warningElement) {
+        if (value > 4) {
+            warningElement.innerHTML = '<span style="color:red">To ensure optimal output quality, please set the <code>Outer Scale</code> to a value of 4 or less. The suggested range is from 1 to 4.</span>';
+        } else {
+            warningElement.innerHTML = '';
+        }
+    }
+    return value; // Return the value to maintain slider functionality
+}

apps/gradio_app/static/styles.css

@@ -0,0 +1,543 @@
+/* @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap'); */
+/* ─── palette ───────────────────────────────────────────── */
+body, .gradio-container {
+  font-family: 'Inter', sans-serif;
+  background: #FFFBF7;
+  color: #0F172A;
+}
+a {
+  color: #F97316;
+  text-decoration: none;
+  font-weight: 600;
+}
+a:hover { color: #C2410C; }
+/* ─── headline ──────────────────────────────────────────── */
+#titlebar {
+  text-align: center;
+  margin-top: 2.4rem;
+  margin-bottom: .9rem;
+}
+/* ─── card look ─────────────────────────────────────────── */
+.gr-block,
+.gr-box,
+.gr-row,
+#cite-wrapper {
+  border: 1px solid #F8C89B;
+  border-radius: 10px;
+  background: #fff;
+  box-shadow: 0 3px 6px rgba(0, 0, 0, .05);
+}
+.gr-gallery-item { background: #fff; }
+/* ─── controls / inputs ─────────────────────────────────── */
+.gr-button-primary,
+#copy-btn {
+  background: linear-gradient(90deg, #F97316 0%, #C2410C 100%);
+  border: none;
+  color: #fff;
+  border-radius: 6px;
+  font-weight: 600;
+  transition: transform .12s ease, box-shadow .12s ease;
+}
+.gr-button-primary:hover,
+#copy-btn:hover {
+  transform: translateY(-2px);
+  box-shadow: 0 4px 12px rgba(249, 115, 22, .35);
+}
+.gr-dropdown input {
+  border: 1px solid #F9731699;
+}
+.preview img,
+.preview canvas { object-fit: contain !important; }
+/* ─── hero section ─────────────────────────────────────── */
+#hero-wrapper { text-align: center; }
+#hero-badge {
+  display: inline-block;
+  padding: .85rem 1.2rem;
+  border-radius: 8px;
+  background: #FFEAD2;
+  border: 1px solid #F9731655;
+  font-size: .95rem;
+  font-weight: 600;
+  margin-bottom: .5rem;
+}
+#hero-links {
+  font-size: .95rem;
+  font-weight: 600;
+  margin-bottom: 1.6rem;
+}
+#hero-links img {
+  height: 22px;
+  vertical-align: middle;
+  margin-left: .55rem;
+}
+/* ─── score area ───────────────────────────────────────── */
+#score-area {
+  text-align: center;
+}
+.title-container {
+    display: flex;
+    align-items: center;
+    gap: 12px;
+    justify-content: center;
+    margin-bottom: 10px;
+    text-align: center;
+}
+.match-badge {
+  display: inline-block;
+  padding: .35rem .9rem;
+  border-radius: 9999px;
+  font-weight: 600;
+  font-size: 1.25rem;
+}
+/* ─── citation card ────────────────────────────────────── */
+#cite-wrapper {
+  position: relative;
+  padding: .9rem 1rem;
+  margin-top: 2rem;
+}
+#cite-wrapper code {
+  font-family: SFMono-Regular, Consolas, monospace;
+  font-size: .84rem;
+  white-space: pre-wrap;
+  color: #0F172A;
+}
+#copy-btn {
+  position: absolute;
+  top: .55rem;
+  right: .6rem;
+  padding: .18rem .7rem;
+  font-size: .72rem;
+  line-height: 1;
+}
+/* ─── dark mode ────────────────────────────────────── */
+.dark body,
+.dark .gradio-container {
+  background-color: #332a22;
+  color: #e5e7eb;
+}
+.dark .gr-block,
+.dark .gr-box,
+.dark .gr-row {
+  background-color: #332a22;
+  border: 1px solid #4b5563;
+}
+.dark .gr-dropdown input {
+  background-color: #332a22;
+  color: #f1f5f9;
+  border: 1px solid #F97316aa;
+}
+.dark #hero-badge {
+  background: #334155;
+  border: 1px solid #F9731655;
+  color: #fefefe;
+}
+.dark #cite-wrapper {
+  background-color: #473f38;
+}
+.dark #bibtex {
+    color: #f8fafc !important;
+}
+.dark .card {
+  background-color: #473f38;
+}
+/* ─── switch logo for light/dark theme ─────────────── */
+.logo-dark { display: none; }
+.dark .logo-light { display: none; }
+.dark .logo-dark { display: inline; }
+
+/* https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap */
+
+/* cyrillic-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2JL7SUc.woff2) format('woff2');
+  unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;
+}
+/* cyrillic */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa0ZL7SUc.woff2) format('woff2');
+  unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;
+}
+/* greek-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2ZL7SUc.woff2) format('woff2');
+  unicode-range: U+1F00-1FFF;
+}
+/* greek */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1pL7SUc.woff2) format('woff2');
+  unicode-range: U+0370-0377, U+037A-037F, U+0384-038A, U+038C, U+038E-03A1, U+03A3-03FF;
+}
+/* vietnamese */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2pL7SUc.woff2) format('woff2');
+  unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;
+}
+/* latin-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa25L7SUc.woff2) format('woff2');
+  unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C0, U+2113, U+2C60-2C7F, U+A720-A7FF;
+}
+/* latin */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 400;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7.woff2) format('woff2');
+  unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+/* cyrillic-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2JL7SUc.woff2) format('woff2');
+  unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;
+}
+/* cyrillic */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa0ZL7SUc.woff2) format('woff2');
+  unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;
+}
+/* greek-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2ZL7SUc.woff2) format('woff2');
+  unicode-range: U+1F00-1FFF;
+}
+/* greek */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1pL7SUc.woff2) format('woff2');
+  unicode-range: U+0370-0377, U+037A-037F, U+0384-038A, U+038C, U+038E-03A1, U+03A3-03FF;
+}
+/* vietnamese */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2pL7SUc.woff2) format('woff2');
+  unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;
+}
+/* latin-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa25L7SUc.woff2) format('woff2');
+  unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C0, U+2113, U+2C60-2C7F, U+A720-A7FF;
+}
+/* latin */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 500;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7.woff2) format('woff2');
+  unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+/* cyrillic-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2JL7SUc.woff2) format('woff2');
+  unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;
+}
+/* cyrillic */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa0ZL7SUc.woff2) format('woff2');
+  unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;
+}
+/* greek-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2ZL7SUc.woff2) format('woff2');
+  unicode-range: U+1F00-1FFF;
+}
+/* greek */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1pL7SUc.woff2) format('woff2');
+  unicode-range: U+0370-0377, U+037A-037F, U+0384-038A, U+038C, U+038E-03A1, U+03A3-03FF;
+}
+/* vietnamese */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2pL7SUc.woff2) format('woff2');
+  unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;
+}
+/* latin-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa25L7SUc.woff2) format('woff2');
+  unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C0, U+2113, U+2C60-2C7F, U+A720-A7FF;
+}
+/* latin */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 600;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7.woff2) format('woff2');
+  unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+/* cyrillic-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2JL7SUc.woff2) format('woff2');
+  unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;
+}
+/* cyrillic */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa0ZL7SUc.woff2) format('woff2');
+  unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;
+}
+/* greek-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2ZL7SUc.woff2) format('woff2');
+  unicode-range: U+1F00-1FFF;
+}
+/* greek */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1pL7SUc.woff2) format('woff2');
+  unicode-range: U+0370-0377, U+037A-037F, U+0384-038A, U+038C, U+038E-03A1, U+03A3-03FF;
+}
+/* vietnamese */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2pL7SUc.woff2) format('woff2');
+  unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;
+}
+/* latin-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa25L7SUc.woff2) format('woff2');
+  unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C0, U+2113, U+2C60-2C7F, U+A720-A7FF;
+}
+/* latin */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 700;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7.woff2) format('woff2');
+  unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+/* cyrillic-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2JL7SUc.woff2) format('woff2');
+  unicode-range: U+0460-052F, U+1C80-1C8A, U+20B4, U+2DE0-2DFF, U+A640-A69F, U+FE2E-FE2F;
+}
+/* cyrillic */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa0ZL7SUc.woff2) format('woff2');
+  unicode-range: U+0301, U+0400-045F, U+0490-0491, U+04B0-04B1, U+2116;
+}
+/* greek-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2ZL7SUc.woff2) format('woff2');
+  unicode-range: U+1F00-1FFF;
+}
+/* greek */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1pL7SUc.woff2) format('woff2');
+  unicode-range: U+0370-0377, U+037A-037F, U+0384-038A, U+038C, U+038E-03A1, U+03A3-03FF;
+}
+/* vietnamese */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa2pL7SUc.woff2) format('woff2');
+  unicode-range: U+0102-0103, U+0110-0111, U+0128-0129, U+0168-0169, U+01A0-01A1, U+01AF-01B0, U+0300-0301, U+0303-0304, U+0308-0309, U+0323, U+0329, U+1EA0-1EF9, U+20AB;
+}
+/* latin-ext */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa25L7SUc.woff2) format('woff2');
+  unicode-range: U+0100-02BA, U+02BD-02C5, U+02C7-02CC, U+02CE-02D7, U+02DD-02FF, U+0304, U+0308, U+0329, U+1D00-1DBF, U+1E00-1E9F, U+1EF2-1EFF, U+2020, U+20A0-20AB, U+20AD-20C0, U+2113, U+2C60-2C7F, U+A720-A7FF;
+}
+/* latin */
+@font-face {
+  font-family: 'Inter';
+  font-style: normal;
+  font-weight: 800;
+  font-display: swap;
+  src: url(https://fonts.gstatic.com/s/inter/v19/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7.woff2) format('woff2');
+  unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+
+/* title_css */
+#title {
+  font-size: 2.6rem;
+  font-weight: 800;
+  margin: 0;
+  line-height: 1.25;
+  color: #0F172A;
+}
+/* brand class is passed in title parameter */
+#title .brand {
+  background: linear-gradient(90deg, #F97316 0%, #C2410C 90%);
+  -webkit-background-clip: text;
+  color: transparent;
+}
+.dark #title {
+  color: #f8fafc;
+}
+.title-container {
+    display: flex;
+    align-items: center;
+    gap: 12px;
+    justify-content: center;
+    margin-bottom: 10px;
+    text-align: center;
+}
+
+/* Dark Mode */
+@media (prefers-color-scheme: dark) {
+  body { @extend .dark; }
+}
+/* Smaller size for input image */
+.input-image img {
+    max-width: 300px; 
+    height: auto;
+}
+/* Larger size for output image */
+.output-image img {
+    max-width: 500px; 
+    height: auto;
+}
+
+/* Add styling for warning message */
+.warning-message {
+    color: red;
+    font-size: 14px;
+    margin-top: 5px;
+    display: block;
+}
+#warning-text {
+    min-height: 20px; /* Ensure space for warning */
+}
+/*Components for Gradio App*/
+.quote-container {
+    border-left: 5px solid #007bff;
+    padding-left: 15px;
+    margin-bottom: 15px;
+    font-style: italic;
+}
+.attribution p {
+    margin: 10px 0;
+}
+.badge {
+    display: inline-block;
+    border-radius: 4px;
+    text-decoration: none;
+    font-size: 14px;
+    transition: background-color 0.3s;
+}
+.badge:hover {
+    background-color: #0056b3;
+}
+.badge img {
+    vertical-align: middle;
+    margin-right: 5px;
+}
+.source {   
+    font-size: 14px;
+}

apps/old-gradio_app.py

@@ -0,0 +1,136 @@
+import gradio as gr
+import os
+import sys
+import json
+from PIL import Image
+
+from gradio_app.project_info import (
+    CONTENT_DESCRIPTION,
+    CONTENT_IN_1, CONTENT_IN_2,
+    CONTENT_OUT_1, CONTENT_OUT_2
+)
+# Add the project root directory to the Python path
+sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
+
+from src.anime_super_resolution.infer import infer
+
+def run_inference(input_image, model_id, outer_scale):
+    if not input_image:
+        return None, "Please upload an image."
+    
+    models_config_path = "configs/model_ckpts.yaml"
+    try:
+        output_image = infer(
+            input_path=input_image,
+            model_id=model_id,
+            models_config_path=models_config_path,
+            outer_scale=outer_scale,
+        )
+        return output_image, "Inference completed successfully!"
+    except Exception as e:
+        return None, f"Error during inference: {str(e)}"
+
+def load_examples():
+    examples = []
+    examples_base_path = os.path.join("apps", "assets", "examples", "Real-ESRGAN-Anime-finetuning")
+    
+    for folder in ["1", "2", "3", "4"]:
+        folder_path = os.path.join(examples_base_path, folder)
+        config_path = os.path.join(folder_path, "config.json")
+        
+        if os.path.exists(config_path):
+            with open(config_path, 'r') as f:
+                config = json.load(f)
+                input_filename = config.get("input_file", "input.jpg")
+                output_filename = config.get("output_file", "output.jpg")
+                outer_scale = config.get("outer_scale", 4)
+                
+                input_image_path = os.path.join(folder_path, input_filename)
+                output_image_path = os.path.join(folder_path, output_filename)
+                
+                if os.path.exists(input_image_path) and os.path.exists(output_image_path):
+                    input_image_data = Image.open(input_image_path)
+                    output_image_data = Image.open(output_image_path)
+                    examples.append([input_image_data, output_image_data, outer_scale])
+    
+    return examples
+
+def select_example(evt: gr.SelectData, examples_data):
+    example_index = evt.index
+    input_image_data, output_image_data, outer_scale = examples_data[example_index]
+    return input_image_data, outer_scale, output_image_data, f"Loaded example with Outer Scale: {outer_scale}"
+
+# Load custom CSS
+custom_css = open("apps/gradio_app/static/styles.css").read()
+
+# JavaScript function to update warning_text Markdown component
+outer_scale_warning = open("apps/gradio_app/static/outer_scale_warning.js").read()
+
+# Define Gradio interface
+with gr.Blocks(css=custom_css) as demo:
+    gr.Markdown("# Anime Super Resolution 🖼️")
+    gr.Markdown(CONTENT_DESCRIPTION)
+    gr.Markdown(CONTENT_IN_1)
+    gr.HTML(CONTENT_IN_2)
+    with gr.Row():
+        with gr.Column(scale=2):
+            input_image = gr.Image(
+                type="filepath",
+                label="Input Image",
+                elem_classes="input-image"
+            )
+            model_id = gr.Textbox(
+                label="Model ID",
+                value="danhtran2mind/Real-ESRGAN-Anime-finetuning"
+            )
+            
+            outer_scale = gr.Slider(
+                minimum=1,
+                maximum=8,
+                step=1,
+                value=2,
+                label="Outer Scale",
+                elem_id="outer-scale-slider"
+            )
+            warning_text = gr.HTML(elem_id="warning-text")  # Added elem_id for JS targeting
+            gr.Markdown(
+                "**Note:** For optimal output quality, set `Outer Scale` to a value between 1 and 4. "
+                "**Values greater than 4 are not recommended**. "
+                "Please ensure `Outer Scale` is greater than or equal to `Inner Scale` (default: 4)."
+            )
+            
+            examples_data = load_examples()
+            submit_button = gr.Button("Run Inference")
+        
+        with gr.Column(scale=3):
+            output_image = gr.Image(
+                label="Output Image",
+                elem_classes="output-image"
+            )
+            output_text = gr.Textbox(label="Status")
+    
+    # Client-side warning update for warning_text
+    outer_scale.change(
+        fn=lambda x: x,  # Pass-through function to satisfy Gradio
+        inputs=outer_scale,
+        outputs=outer_scale,  # Return value to maintain slider functionality
+        js=outer_scale_warning  # Update warning_text via JavaScript
+    )
+    
+    gr.Examples(
+        examples=[[input_img, output_img, outer_scale] for input_img, output_img, outer_scale in examples_data],
+        inputs=[input_image, output_image, outer_scale],
+        label="Example Inputs",
+        examples_per_page=4
+    )
+        
+    submit_button.click(
+        fn=run_inference,
+        inputs=[input_image, model_id, outer_scale],
+        outputs=[output_image, output_text]
+    )
+    gr.HTML(CONTENT_OUT_1)
+    gr.Markdown(CONTENT_OUT_2)
+
+if __name__ == "__main__":
+    demo.launch(share=True)  # Changed to local launch for safety; use share=True for public URL if needed

assets/examples/Real-ESRGAN-Anime-finetuning/1/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_2T3.png",
+    "output_file": "frame_0001_2T3_real_esrgan_anime_x2.png",
+    "outer_scale": 2
+}

assets/examples/Real-ESRGAN-Anime-finetuning/2/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_9T3.png",
+    "output_file": "frame_0001_9T3_real_esrgan_anime_x4.png",
+    "outer_scale": 4
+}

assets/examples/Real-ESRGAN-Anime-finetuning/3/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_10T3.png",
+    "output_file": "frame_0001_10T3_real_esrgan_anime_x6.png",
+    "outer_scale": 6
+}

assets/examples/Real-ESRGAN-Anime-finetuning/4/config.json

@@ -0,0 +1,5 @@
+{
+    "input_file": "frame_0001_14T3.png",
+    "output_file": "frame_0001_14T3_real_esrgan_anime_x8.png",
+    "outer_scale": 8
+}

configs/Real-ESRGAN-Anime-finetuning.yml

@@ -0,0 +1,188 @@
+# general settings
+name: Real-ESRGAN-Anime-finetuning
+model_type: RealESRGANModel
+scale: 4
+num_gpu: auto
+manual_seed: 0
+
+# ----------------- options for synthesizing training data in RealESRGANModel ----------------- #
+# USM the ground-truth
+l1_gt_usm: True
+percep_gt_usm: True
+gan_gt_usm: False
+
+# the first degradation process
+resize_prob: [0.2, 0.7, 0.1]  # up, down, keep
+resize_range: [0.15, 1.5]
+gaussian_noise_prob: 0.5
+noise_range: [1, 30]
+poisson_scale_range: [0.05, 3]
+gray_noise_prob: 0.4
+jpeg_range: [30, 95]
+
+# the second degradation process
+second_blur_prob: 0.8
+resize_prob2: [0.3, 0.4, 0.3]  # up, down, keep
+resize_range2: [0.3, 1.2]
+gaussian_noise_prob2: 0.5
+noise_range2: [1, 25]
+poisson_scale_range2: [0.05, 2.5]
+gray_noise_prob2: 0.4
+jpeg_range2: [30, 95]
+
+gt_size: 256
+queue_size: 180
+
+# dataset and data loader settings
+datasets:
+  train:
+    name: DF2K+OST
+    type: RealESRGANDataset
+    dataroot_gt: './data'
+    meta_info: './data/meta_info/meta_info_multiscale.txt'
+    io_backend:
+      type: disk
+
+    blur_kernel_size: 21
+    kernel_list: ['iso', 'aniso', 'generalized_iso', 'generalized_aniso', 'plateau_iso', 'plateau_aniso']
+    kernel_prob: [0.45, 0.25, 0.12, 0.03, 0.12, 0.03]
+    sinc_prob: 0.1
+    blur_sigma: [0.2, 3]
+    betag_range: [0.5, 4]
+    betap_range: [1, 2]
+
+    blur_kernel_size2: 21
+    kernel_list2: ['iso', 'aniso', 'generalized_iso', 'generalized_aniso', 'plateau_iso', 'plateau_aniso']
+    kernel_prob2: [0.45, 0.25, 0.12, 0.03, 0.12, 0.03]
+    sinc_prob2: 0.1
+    blur_sigma2: [0.2, 1.5]
+    betag_range2: [0.5, 4]
+    betap_range2: [1, 2]
+
+    final_sinc_prob: 0.8
+
+    gt_size: 256
+    use_hflip: True
+    use_rot: False
+
+    # data loader
+    use_shuffle: true
+    num_worker_per_gpu: 5
+    batch_size_per_gpu: 10
+    dataset_enlarge_ratio: 1
+    prefetch_mode: ~
+
+  # Uncomment these for validation
+  # val:
+  #   name: validation
+  #   type: PairedImageDataset
+  #   dataroot_gt: path_to_gt
+  #   dataroot_lq: path_to_lq
+  #   io_backend:
+  #     type: disk
+
+# network structures
+network_g:
+  type: RRDBNet
+  num_in_ch: 3
+  num_out_ch: 3
+  num_feat: 64
+  num_block: 23
+  num_grow_ch: 32
+
+network_d:
+  type: UNetDiscriminatorSN
+  num_in_ch: 3
+  num_feat: 64
+  skip_connection: True
+
+# path
+path:
+  # use the pre-trained Real-ESRNet model
+  pretrain_network_g: ./ckpts/Real-ESRGAN/RealESRGAN_x4plus.pth
+  param_key_g: params_ema
+  strict_load_g: true
+  pretrain_network_d: ./ckpts/Real-ESRGAN/RealESRGAN_x4plus_netD.pth
+  param_key_d: params
+  strict_load_d: true
+  resume_state: ~
+
+# training settings
+train:
+  ema_decay: 0.999
+  optim_g:
+    type: Adam
+    lr: !!float 1e-4
+    weight_decay: 0
+    betas: [0.9, 0.99]
+  optim_d:
+    type: Adam
+    lr: !!float 1e-4
+    weight_decay: 0
+    betas: [0.9, 0.99]
+
+  scheduler:
+    type: MultiStepLR
+    milestones: [1]
+    gamma: 0.5
+
+  total_iter: 20000
+  warmup_iter: -1  # no warm up
+
+  # losses
+  pixel_opt:
+    type: L1Loss
+    loss_weight: 1.0
+    reduction: mean
+  # perceptual loss (content and style losses)
+  perceptual_opt:
+    type: PerceptualLoss
+    layer_weights:
+      # before relu
+      'conv1_2': 0.1
+      'conv2_2': 0.1
+      'conv3_4': 1
+      'conv4_4': 1
+      'conv5_4': 1
+    vgg_type: vgg19
+    use_input_norm: true
+    perceptual_weight: !!float 1.0
+    style_weight: 0
+    range_norm: false
+    criterion: l1
+  # gan loss
+  gan_opt:
+    type: GANLoss
+    gan_type: vanilla
+    real_label_val: 1.0
+    fake_label_val: 0.0
+    loss_weight: !!float 1e-1
+
+  net_d_iters: 1
+  net_d_init_iters: 0
+
+# Uncomment these for validation
+# validation settings
+# val:
+#   val_freq: !!float 5e3
+#   save_img: True
+
+#   metrics:
+#     psnr: # metric name
+#       type: calculate_psnr
+#       crop_border: 4
+#       test_y_channel: false
+
+# logging settings
+logger:
+  print_freq: 100
+  save_checkpoint_freq: !!float 5e3
+  use_tb_logger: true
+  wandb:
+    project: ~
+    resume_id: ~
+
+# dist training settings
+dist_params:
+  backend: nccl
+  port: 29500

configs/model_ckpts.yaml

@@ -0,0 +1,22 @@
+- model_id: "danhtran2mind/Real-ESRGAN-Anime-finetuning"
+  local_dir: "ckpts/Real-ESRGAN-Anime-finetuning"
+  filename: "net_g_latest.pth"
+  for_inference: true
+  base_model: false
+  platform: "HuggingFace"
+
+- model_id: "xinntao/Real-ESRGAN-x4plus-netD"
+  url: "https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.2.3/RealESRGAN_x4plus_netD.pth"
+  local_dir: "ckpts/Real-ESRGAN"
+  filename: "RealESRGAN_x4plus_netD.pth"
+  for_inference: false
+  base_model: true
+  platform: "GitHub"
+
+- model_id: "xinntao/Real-ESRGAN-x4plus"
+  url: "https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pth"
+  local_dir: "ckpts/Real-ESRGAN"
+  filename: "RealESRGAN_x4plus.pth"
+  for_inference: true
+  base_model: true
+  platform: "GitHub"

data/.gitignore

@@ -0,0 +1,2 @@
+*
+!.gitignore

docs/inference/inference_doc.md

@@ -0,0 +1,116 @@
+# Inference Documentation
+
+This document provides instructions for using the Real-ESRGAN inference script to perform super-resolution on anime images. The script uses a pre-trained Real-ESRGAN model to upscale images, with configurable input and output options.
+
+## Prerequisites
+
+- **Python Libraries**: Ensure the following Python packages are installed:
+  - `argparse`
+  - `PIL` (Pillow)
+  - `numpy`
+  - `torch`
+  - `opencv-python` (cv2)
+  - `pyyaml`
+  - `huggingface_hub`
+- **Model Configuration**: A YAML file specifying model details (model ID, local directory, and filename).
+- **Input Image**: A valid image file (e.g., PNG, JPEG) in RGB format.
+- **Hardware**: CUDA-compatible GPU (optional, for faster processing) or CPU.
+
+## Script Overview
+
+The script (`inference.py`) performs super-resolution on an input image using the Real-ESRGAN model. It supports:
+- Downloading model weights from Hugging Face if not available locally.
+- Upscaling images using an inner scale (model-specific) and an optional outer scale (post-processing resizing).
+- Saving the upscaled image to a specified output path or a default location.
+
+## Command-Line Arguments
+
+The script accepts the following command-line arguments:
+
+| Argument                | Type | Required | Default | Description                                                                 |
+|-------------------------|------|----------|---------|-----------------------------------------------------------------------------|
+| `--input_path`          | str  | Yes      | None    | Path to the input image file (e.g., `image.png`).                           |
+| `--output_path`         | str  | No       | None    | Path to save the upscaled image. If not provided, the image is returned but not saved automatically. |
+| `--model_id`            | str  | Yes      | None    | Model ID for the Real-ESRGAN model (e.g., `danhtran2mind/Real-ESRGAN-Anime-finetuning`). |
+| `--models_config_path`  | str  | Yes      | None    | Path to the YAML configuration file containing model details.                |
+| `--batch_size`          | int  | No       | 1       | Batch Ascertain batch size (not used in this implementation).                    |
+| `--outer_scale`        | int  | Yes      | None    | Desired final scale factor for super-resolution (e.g., 4, 8).               |
+| `--inner_scale`        | int  | No       | 4       | Internal scale factor used by the model (typically 4).                      |
+
+## Usage
+
+1. **Prepare the Models Configuration File**:
+   Create a YAML file (e.g., `models_config.yaml`) with the following structure:
+
+   ```yaml
+   - model_id: "danhtran2mind/Real-ESRGAN-Anime-finetuning"
+     local_dir: "./weights"
+     filename: "model.pth"
+   ```
+
+   This file specifies the model ID, local directory for weights, and the filename of the model checkpoint.
+
+2. **Run the Script**:
+   Use the following command to run the inference:
+
+   ```bash
+   python inference.py \
+      --input_path path/to/input/image.png \
+      --output_path path/to/output/image.png \
+      --model_id danhtran2mind/Real-ESRGAN-Anime-finetuning \
+      --models_config_path path/to/models_config.yaml \
+      --outer_scale 4
+   ```
+
+   Example:
+
+   ```bash
+   python inference.py \
+      --input_path input.png \
+      --output_path output.png \
+      --model_id danhtran2mind/Real-ESRGAN-Anime-finetuning \
+      --models_config_path models_config.yaml \
+      --outer_scale 8
+   ```
+
+3. **Output**:
+   - The script processes the input image and applies super-resolution.
+   - If `--output_path` is provided, the upscaled image is saved to the specified path.
+   - If `--outer_scale` differs from `--inner_scale`, the output image is resized using OpenCV's `INTER_CUBIC` (for upscaling) or `INTER_AREA` (for downscaling) interpolation.
+
+## How It Works
+
+1. **Model Loading**:
+   - The script reads the `models_config_path` YAML file to locate the model configuration.
+   - If the model weights are not found locally, they are downloaded from the Hugging Face Hub using the specified `model_id` and `filename`.
+   - The Real-ESRGAN model is initialized with the specified `inner_scale` and loaded with the weights.
+
+2. **Image Processing**:
+   - The input image is opened and converted to RGB format using Pillow.
+   - The Real-ESRGAN model upscales the image by the `inner_scale` factor.
+   - If `outer_scale` differs from `inner_scale`, the image is further resized to achieve the desired scale using OpenCV.
+
+3. **Output Handling**:
+   - The upscaled image is saved to `output_path` if provided.
+   - The processed image is returned as a Pillow Image object.
+
+## Notes
+
+- **Device Selection**: The script automatically uses CUDA if available; otherwise, it falls back to CPU.
+- **Model Weights**: Ensure the `local_dir` specified in the YAML file exists or is writable for downloading weights.
+- **Outer vs. Inner Scale**:
+  - `inner_scale` is the scale factor used by the Real-ESRGAN model (typically fixed at 4).
+  - `outer_scale` is the final desired scale, achieved through additional resizing if necessary.
+- **Batch Size**: The `--batch_size` argument is included but not used in this implementation, as the script processes one image at a time.
+
+## Example Models Configuration File
+
+Here is an example `models_config.yaml`:
+
+<xaiArtifact artifact_id="0b60a214-8c91-48ed-ad50-fae3467a0508" artifact_version_id="35ae4a0a-da96-44d0-b8ed-7c1d62b59527" title="models_config.yaml" contentType="text/yaml">
+
+```yaml
+- model_id: "danhtran2mind/Real-ESRGAN-Anime-finetuning"
+  local_dir: "./weights"
+  filename: "model.pth"
+```