Merge pull request #1 from naseemap47/app

App

.github/workflows/docker-image.yml

@@ -0,0 +1,33 @@
+name: Docker Image CI
+
+on:
+  push:
+    branches: [ "master" ]
+  pull_request:
+    branches: [ "master" ]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+      -
+        name: Checkout
+        uses: actions/checkout@v3
+      -
+        name: Login to Docker Hub
+        uses: docker/login-action@v2
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+      -
+        name: Build and push
+        uses: docker/build-push-action@v4
+        with:
+          context: .
+          file: ./Dockerfile
+          push: true
+          tags: ${{ secrets.DOCKERHUB_USERNAME }}/streamlit-nafnet:latest

.gitignore

@@ -158,3 +158,6 @@ cython_debug/
 #  and can be added to the global gitignore or merged into this file.  For a more nuclear
 #  option (not recommended) you can uncomment the following to ignore the entire idea folder.
 #.idea/
+
+# Model
+*.pth

Dockerfile

@@ -0,0 +1,18 @@
+FROM ubuntu:20.04
+ARG DEBIAN_FRONTEND=noninteractive
+COPY . /home
+RUN apt-get update && \
+    apt-get install -y \
+    python3 \
+    python3-pip \
+    ffmpeg \
+    libsm6 \
+    libxext6
+WORKDIR /home
+RUN pip install torch torchvision torchaudio
+RUN pip install -r requirements.txt
+RUN pip install --upgrade --no-cache-dir gdown
+RUN python3 setup.py develop --no_cuda_ext
+RUN pip install streamlit
+RUN python3 -c "import gdown;gdown.download('https://drive.google.com/uc?id=14D4V4raNYIOhETfcuuLI3bGLB-OYIv6X', './experiments/pretrained_models/', quiet=False)"
+CMD [ "streamlit", "run", "app.py" ]

README.md

@@ -1,2 +1,25 @@
 # ImageRestoration-Streamlit
+[<img src="https://img.shields.io/badge/Docker-Image-blue.svg?logo=docker">](<https://hub.docker.com/repository/docker/naseemap47/streamlit-nafnet>) <br>
 Restore image using NAFNet model with streamlit dashboard
+
+## Installtion
+1. **Clone this Repo**
+```
+git clone https://github.com/naseemap47/ImageRestoration-Streamlit.git
+cd ImageRestoration-Streamlit
+```
+2. **Install Libraries**
+```
+pip install -r requirements.txt
+pip install --upgrade --no-cache-dir gdown
+python3 setup.py develop --no_cuda_ext
+pip install streamlit
+```
+## Download pretrained models
+```
+python3 -c "import gdown;gdown.download('https://drive.google.com/uc?id=14D4V4raNYIOhETfcuuLI3bGLB-OYIv6X', './experiments/pretrained_models/', quiet=False)"
+```
+## Run Dashboard
+```
+streamlit run app.py
+```

app.py

@@ -0,0 +1,53 @@
+import streamlit as st
+from basicsr.models import create_model
+from basicsr.utils import img2tensor as _img2tensor, tensor2img
+from basicsr.utils.options import parse
+import numpy as np
+import cv2
+
+
+def img2tensor(img, bgr2rgb=False, float32=True):
+    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+    img = img.astype(np.float32) / 255.
+    return _img2tensor(img, bgr2rgb=bgr2rgb, float32=float32)
+
+def single_image_inference(model, img):
+	model.feed_data(data={'lq': img.unsqueeze(dim=0)})
+
+	if model.opt['val'].get('grids', False):
+		model.grids()
+
+	model.test()
+
+	if model.opt['val'].get('grids', False):
+		model.grids_inverse()
+
+	visuals = model.get_current_visuals()
+	sr_img = tensor2img([visuals['result']])
+	return sr_img
+
+
+opt_path = 'options/test/REDS/NAFNet-width64.yml'
+opt = parse(opt_path, is_train=False)
+opt['dist'] = False
+NAFNet = create_model(opt)
+
+################## APP ##################
+st.title('Unblur Image - NAFNet')
+
+# Upload image
+upload_img_file = st.file_uploader(
+    'Upload Image', type=['jpg', 'jpeg', 'png'])
+FRAME_WINDOW = st.image([], channels='BGR')
+
+if upload_img_file is not None:
+    file_bytes = np.asarray(
+        bytearray(upload_img_file.read()), dtype=np.uint8)
+    img_input = cv2.imdecode(file_bytes, 1)
+    FRAME_WINDOW.image(img_input, channels='BGR')
+
+if st.button('Submit', use_container_width=True):
+    st.subheader('Result')
+    inp = img2tensor(img_input)
+    out_img = single_image_inference(NAFNet, inp)
+    st.image(out_img, channels='BGR')