Linted files

.gitignore

@@ -9,4 +9,4 @@ cv4e_lecture13/datasets/
 coverage/
 
 __pycache__/
-docs/build
+docs/build

.pre-commit-config.yaml

@@ -30,6 +30,7 @@ repos:
     rev: v3.1.0
     hooks:
       - id: check-ast
+      - id: check-toml
       - id: check-executables-have-shebangs
       - id: check-docstring-first
       - id: double-quote-string-fixer

LICENSE

@@ -18,4 +18,4 @@ 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
+SOFTWARE

README.md

@@ -2,8 +2,8 @@
 title: CV4Ecology School - Lecture 13
 metaTitle: "Serving, Hosting, and Deploying Models and Quality Control"
 emoji: 🌎
-colorFrom: red
-colorTo: purple
+colorFrom: blue
+colorTo: green
 sdk: gradio
 sdk_version: 3.1.4
 app_file: app.py

README.rst

@@ -38,7 +38,7 @@ You need to first install Anaconda on your machine.  Below are the instructions
 Once Anaconda is installed, you will need an environment and the following packages installed
 
 .. code:: bash
-   
+
    # Create Environment
    conda create --name cv4e
    conda activate cv4e
@@ -46,16 +46,16 @@ Once Anaconda is installed, you will need an environment and the following packa
    # Install Python dependencies
    conda install pip
 
-   conda install -r requirements.txt 
+   conda install -r requirements.txt
    conda install pytorch torchvision -c pytorch-nightly
 
 How to Run
 ----------
 
-The lecture materials will run as a single executable.  The MNIST dataset must be downloaded from the internet for this script to run correctly, so Internet access is required at first to download the files once.  It is recommended to use `ipython` and to copy sections of code into and inspecting the 
+The lecture materials will run as a single executable.  The MNIST dataset must be downloaded from the internet for this script to run correctly, so Internet access is required at first to download the files once.  It is recommended to use `ipython` and to copy sections of code into and inspecting the
 
 .. code:: bash
-   
+
    # Run with Python
    python lecture.py
 

app.py

@@ -1,9 +1,10 @@
-from cv4e_lecture13 import model, utils
+# -*- coding: utf-8 -*-
+import gradio as gr
 import torch
 from PIL import Image, ImageOps  # NOQA
 from torchvision.transforms import Compose, Resize, ToTensor
-import gradio as gr
 
+from cv4e_lecture13 import model, utils
 
 config = 'cv4e_lecture13/configs/mnist_resnet18.yaml'
 
@@ -11,7 +12,7 @@ log = utils.init_logging()
 cfg = utils.init_config(config, log)
 device = cfg.get('device')
 
-cfg['output'] = 'cv4e_lecture13/%s' % (cfg['output'], )
+cfg['output'] = 'cv4e_lecture13/{}'.format(cfg['output'])
 
 net, _, _ = model.load(cfg)
 net.eval()
@@ -20,10 +21,7 @@ net.eval()
 def predict(inp):
     inp = ImageOps.grayscale(inp)
 
-    transforms = Compose([             
-        Resize((cfg['image_size'])), 
-        ToTensor()                   
-    ])
+    transforms = Compose([Resize(cfg['image_size']), ToTensor()])
     inp = transforms(inp).unsqueeze(0)
     data = inp.to(device)
 
@@ -33,7 +31,10 @@ def predict(inp):
     confidences = torch.softmax(prediction[0], dim=0).cpu().numpy()
     confidences = list(enumerate(confidences))
     confidences = [
-        (str(label), float(conf) , )
+        (
+            str(label),
+            float(conf),
+        )
         for label, conf in confidences
     ]
     confidences = dict(confidences)
@@ -42,13 +43,10 @@ def predict(inp):
 
 
 interface = gr.Interface(
-    fn=predict, 
+    fn=predict,
     inputs=gr.Image(type='pil'),
     outputs=gr.Label(num_top_classes=3),
-    examples=[
-        f'examples/example_{index}.jpg'
-        for index in range(1, 31)
-    ]
+    examples=[f'examples/example_{index}.jpg' for index in range(1, 31)],
 )
 
 interface.launch()

cv4e_lecture13/__init__.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 '''
     2022 Benjamin Kellenberger
 '''

cv4e_lecture13/configs/mnist_resnet18.yaml

@@ -18,4 +18,4 @@ learning_rate: 0.001
 weight_decay: 0.001
 
 # output params
-output: models
+output: models

cv4e_lecture13/dataset.py

@@ -1,13 +1,15 @@
+# -*- coding: utf-8 -*-
 '''
     Model implementation.
     We'll be using a "simple" ResNet-18 for image classification here.
 
     2022 Benjamin Kellenberger
 '''
+from os.path import abspath
+
 import torch
 from torchvision import datasets
 from torchvision.transforms import Compose, Resize, ToTensor
-from os.path import abspath
 
 
 def load(cfg):
@@ -20,32 +22,26 @@ def load(cfg):
 
     train = torch.utils.data.DataLoader(
         datasets.MNIST(
-            root, 
-            train=True, 
-            transform=Compose([             
-                Resize((cfg['image_size'])), 
-                ToTensor()                   
-            ]),
-            download=True
-        ), 
-        batch_size=cfg.get('batch_size'), 
-        shuffle=True, 
-        num_workers=cfg.get('num_workers')
+            root,
+            train=True,
+            transform=Compose([Resize(cfg['image_size']), ToTensor()]),
+            download=True,
+        ),
+        batch_size=cfg.get('batch_size'),
+        shuffle=True,
+        num_workers=cfg.get('num_workers'),
     )
-    
+
     test = torch.utils.data.DataLoader(
         datasets.MNIST(
-            root, 
-            train=False, 
-            transform=Compose([             
-                Resize((cfg['image_size'])), 
-                ToTensor()                   
-            ]),
-            download=True
-        ), 
-        batch_size=cfg.get('batch_size'), 
-        shuffle=False, 
-        num_workers=cfg.get('num_workers')
+            root,
+            train=False,
+            transform=Compose([Resize(cfg['image_size']), ToTensor()]),
+            download=True,
+        ),
+        batch_size=cfg.get('batch_size'),
+        shuffle=False,
+        num_workers=cfg.get('num_workers'),
     )
 
     return train, test

cv4e_lecture13/model.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 '''
     Model implementation.
     We'll be using a "simple" ResNet-18 for image classification here.
@@ -5,17 +6,17 @@
     2022 Benjamin Kellenberger
 '''
 
+import glob
+import os
+from os.path import exists, split, splitext
+
+import numpy as np
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-import numpy as np
-import glob
-import os
-from os.path import split, splitext, exists
 
 
 class SmallModel(nn.Module):
-
     @classmethod
     def load(cls, cfg):
         log = cfg.get('log')
@@ -30,11 +31,7 @@ class SmallModel(nn.Module):
         filepaths = sorted(glob.glob(f'{output}/*.pt'))
 
         if len(filepaths) > 1:
-            filepaths = [
-                filepath 
-                for filepath in filepaths 
-                if 'best.pt' not in filepath
-            ]
+            filepaths = [filepath for filepath in filepaths if 'best.pt' not in filepath]
 
         if len(filepaths):
             filepath = filepaths[-1]
@@ -49,7 +46,7 @@ class SmallModel(nn.Module):
                 epoch = int(splitext(filename)[0])
             except ValueError:
                 pass
-                
+
             filepath = f'{output}/best.pt'
             if exists(filepath):
                 state = torch.load(open(filepath, 'rb'), map_location='cpu')
@@ -73,7 +70,7 @@ class SmallModel(nn.Module):
     def forward(self, x):
         x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2))
         x = F.max_pool2d(F.relu(self.conv2(x)), 2)
-        x = torch.flatten(x, 1) 
+        x = torch.flatten(x, 1)
         x = F.relu(self.fc1(x))
         x = F.relu(self.fc2(x))
         x = self.fc3(x)

cv4e_lecture13/train.py

@@ -3,20 +3,20 @@
 """
 The lecture materials for Lecture 1: Dataset Prototyping and Visualization
 """
-from . import model, dataset, utils
-import torch
 import click
+import torch
 import torch.nn as nn
-from tqdm import trange
 from torch.optim import Adam
+from tqdm import trange
 
+from . import dataset, model, utils
 
 log = None
 
 
 def inference(cfg, dataloader, net, optimizer, criterion, update):
     '''
-        Our actual training function.
+    Our actual training function.
     '''
     device = cfg.get('device')
 
@@ -44,10 +44,14 @@ def inference(cfg, dataloader, net, optimizer, criterion, update):
         label_ = torch.argmax(prediction, dim=1)
         accuracy += torch.mean((label_ == labels).float()).item()
 
-        prog.set_description('[{:s}] Loss: {:.2f}; Acc: {:.2f}%'.format(type_str, loss / (index + 1), 100.0 * accuracy / (index + 1)))
+        prog.set_description(
+            '[{:s}] Loss: {:.2f}; Acc: {:.2f}%'.format(
+                type_str, loss / (index + 1), 100.0 * accuracy / (index + 1)
+            )
+        )
         prog.update(1)
     prog.close()
-    
+
     loss /= total
     accuracy /= total
 
@@ -55,7 +59,9 @@ def inference(cfg, dataloader, net, optimizer, criterion, update):
 
 
 @click.command()
-@click.option('--config', help='Path to config file', default='configs/mnist_resnet18.yaml')
+@click.option(
+    '--config', help='Path to config file', default='configs/mnist_resnet18.yaml'
+)
 def lecture(config):
     """
     Main function for Lecture 1: Dataset Prototyping and Visualization
@@ -86,15 +92,19 @@ def lecture(config):
     while epoch < epochs:
         log.info(f'Epoch {epoch}/{epochs}')
 
-        loss_train, accuracy_train = inference(cfg, train, net, optimizer, criterion, update=True)
-        loss_test, accuracy_test = inference(cfg, test, net, optimizer, criterion, update=False)
+        loss_train, accuracy_train = inference(
+            cfg, train, net, optimizer, criterion, update=True
+        )
+        loss_test, accuracy_test = inference(
+            cfg, test, net, optimizer, criterion, update=False
+        )
 
         # combine stats and save
         stats = {
             'loss_train': loss_train,
             'loss_val': loss_test,
             'accuracy_train': accuracy_train,
-            'accuracy_test': accuracy_test
+            'accuracy_test': accuracy_test,
         }
 
         best = loss_test < best_loss

cv4e_lecture13/utils.py

@@ -1,16 +1,17 @@
+# -*- coding: utf-8 -*-
 '''
     Various utility functions used (possibly) across scripts.
 
     2022 Benjamin Kellenberger
 '''
 
-import random
-import torch
-from torch.backends import cudnn
 import logging
+import random
 from logging.handlers import TimedRotatingFileHandler
-import yaml
 
+import torch
+import yaml
+from torch.backends import cudnn
 
 DAYS = 21
 
@@ -102,7 +103,9 @@ def init_config(config, log):
         elif torch.backends.mps.is_available():
             cfg['device'] = 'mps'
         else:
-            log.warning(f'WARNING: device set to "{device}" but not available; falling back to CPU...')
+            log.warning(
+                f'WARNING: device set to "{device}" but not available; falling back to CPU...'
+            )
             cfg['device'] = 'cpu'
 
     device = cfg.get('device')

docs/conf.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 # Configuration file for the Sphinx documentation builder.
 #
 # This file only contains a selection of the most common options. For a full
@@ -17,47 +18,47 @@
 
 # -- Project information -----------------------------------------------------
 
-project = "CV4Ecology School, Lecture 13"
-copyright = "2022"
-author = "CV4EcologySchool"
+project = 'CV4Ecology School, Lecture 13'
+copyright = '2022'
+author = 'CV4EcologySchool'
 
 
 # -- General configuration ---------------------------------------------------
 # -- General configuration
 
 extensions = [
-    "sphinx.ext.duration",
-    "sphinx.ext.doctest",
-    "sphinx.ext.autodoc",
-    "sphinx.ext.autosummary",
-    "sphinx.ext.intersphinx",
+    'sphinx.ext.duration',
+    'sphinx.ext.doctest',
+    'sphinx.ext.autodoc',
+    'sphinx.ext.autosummary',
+    'sphinx.ext.intersphinx',
 ]
 
 intersphinx_mapping = {
-    "rtd": ("https://docs.readthedocs.io/en/stable/", None),
-    "python": ("https://docs.python.org/3/", None),
-    "sphinx": ("https://www.sphinx-doc.org/en/master/", None),
+    'rtd': ('https://docs.readthedocs.io/en/stable/', None),
+    'python': ('https://docs.python.org/3/', None),
+    'sphinx': ('https://www.sphinx-doc.org/en/master/', None),
 }
-intersphinx_disabled_domains = ["std"]
+intersphinx_disabled_domains = ['std']
 
-templates_path = ["_templates"]
+templates_path = ['_templates']
 
 # -- Options for EPUB output
-epub_show_urls = "footnote"
+epub_show_urls = 'footnote'
 
 # List of patterns, relative to source directory, that match files and
 # directories to ignore when looking for source files.
 # This pattern also affects html_static_path and html_extra_path.
-exclude_patterns = ["_build", "Thumbs.db", ".DS_Store"]
+exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
 
 # -- Options for HTML output -------------------------------------------------
 
 # The theme to use for HTML and HTML Help pages.  See the documentation for
 # a list of builtin themes.
 #
-html_theme = "sphinx_rtd_theme"
+html_theme = 'sphinx_rtd_theme'
 
 # Add any paths that contain custom static files (such as style sheets) here,
 # relative to this directory. They are copied after the builtin static files,
 # so a file named "default.css" will overwrite the builtin "default.css".
-html_static_path = ["_static"]
+html_static_path = ['_static']

requirements.txt

@@ -4,13 +4,13 @@ Pillow
 numpy
 PyYAML
 tqdm
-pre-commit 
+pre-commit
 cryptography
 argparse
 gradio
-numpy 
-ipython 
-rich 
+numpy
+ipython
+rich
 click
 brunette
 pytest

setup.cfg

@@ -8,7 +8,7 @@ author = CV4EcologySchool
 author_email = cv4ecology@caltech.edu,
 license = MIT
 license_file = LICENSE
-project_urls = 
+project_urls =
     Documentation = https://cv4e_lecture13.readthedocs.io
     Source = https://github.com/CV4EcologySchool
 

setup.py

@@ -2,5 +2,5 @@
 # -*- coding: utf-8 -*-
 import setuptools
 
-if __name__ == "__main__":
+if __name__ == '__main__':
     setuptools.setup()

tests/__init__.py

@@ -1 +0,0 @@
-# -*- coding: utf-8 -*-

tests/conftest.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 import logging
+
 import pytest
 
 log = logging.getLogger('pytest.conftest')  # pylint: disable=invalid-name
@@ -13,10 +14,11 @@ def config():
 @pytest.fixture()
 def cfg(config):
     from cv4e_lecture13 import utils
+
     log = utils.init_logging()
     cfg = utils.init_config(config, log)
 
-    cfg['output'] = 'cv4e_lecture13/%s' % (cfg['output'], )
+    cfg['output'] = 'cv4e_lecture13/{}'.format(cfg['output'])
 
     return cfg
 
@@ -31,6 +33,7 @@ def device(cfg):
 @pytest.fixture()
 def net(cfg):
     from cv4e_lecture13 import model
+
     net, _, _ = model.load(cfg)
     net.eval()
 

tests/test_model.py

@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
-from torchvision.transforms import Compose, Resize, ToTensor
-from PIL import Image, ImageOps
 import torch
+from PIL import Image, ImageOps
+from torchvision.transforms import Compose, Resize, ToTensor
 
 
 def test_architecture_params(net):
@@ -14,10 +14,7 @@ def test_model_prediction(cfg, device, net):
 
     image = ImageOps.grayscale(image)
 
-    transforms = Compose([             
-        Resize((cfg['image_size'])), 
-        ToTensor()                   
-    ])
+    transforms = Compose([Resize(cfg['image_size']), ToTensor()])
     image = transforms(image).unsqueeze(0)
     data = image.to(device)