iapytoo 0.0.2__tar.gz → 0.0.4.dev3__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (56) hide show
  1. iapytoo-0.0.4.dev3/.github/workflows/python-app.yml +78 -0
  2. iapytoo-0.0.4.dev3/.gitignore +168 -0
  3. {iapytoo-0.0.2/iapytoo.egg-info → iapytoo-0.0.4.dev3}/PKG-INFO +1 -1
  4. iapytoo-0.0.4.dev3/dockers/Dockerfile +13 -0
  5. iapytoo-0.0.4.dev3/dockers/Readme.md +7 -0
  6. iapytoo-0.0.4.dev3/dockers/run-python.sh +9 -0
  7. iapytoo-0.0.4.dev3/examples/config_mnist.json +19 -0
  8. iapytoo-0.0.4.dev3/examples/config_wgan.json +25 -0
  9. iapytoo-0.0.4.dev3/examples/mnist.py +106 -0
  10. iapytoo-0.0.4.dev3/examples/wgan_mnist.py +200 -0
  11. iapytoo-0.0.4.dev3/iapytoo/_version.py +16 -0
  12. iapytoo-0.0.4.dev3/iapytoo/utils/cf/config.json +17 -0
  13. iapytoo-0.0.4.dev3/iapytoo/utils/cf/config_test.json +17 -0
  14. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3/iapytoo.egg-info}/PKG-INFO +1 -1
  15. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo.egg-info/SOURCES.txt +13 -0
  16. iapytoo-0.0.4.dev3/iapytoo.egg-info/top_level.txt +3 -0
  17. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/pyproject.toml +8 -12
  18. iapytoo-0.0.4.dev3/requirements.txt +6 -0
  19. iapytoo-0.0.2/iapytoo.egg-info/top_level.txt +0 -1
  20. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/LICENSE +0 -0
  21. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/MANIFEST.in +0 -0
  22. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/README.md +0 -0
  23. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/__init__.py +0 -0
  24. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/dataset/__init__.py +0 -0
  25. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/dataset/scaling.py +0 -0
  26. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/metrics/__init__.py +0 -0
  27. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/metrics/distrib.py +0 -0
  28. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/metrics/metric.py +0 -0
  29. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/metrics/predefined.py +0 -0
  30. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/predictions/__init__.py +0 -0
  31. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/predictions/feature.py +0 -0
  32. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/predictions/plotters.py +0 -0
  33. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/predictions/predictors.py +0 -0
  34. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/predictions/types.py +0 -0
  35. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/__init__.py +0 -0
  36. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/checkpoint.py +0 -0
  37. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/context.py +0 -0
  38. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/factories.py +0 -0
  39. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/logger.py +0 -0
  40. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/loss.py +0 -0
  41. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/training.py +0 -0
  42. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/train/wgan.py +0 -0
  43. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/__init__.py +0 -0
  44. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/arguments.py +0 -0
  45. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/config.py +0 -0
  46. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/display.py +0 -0
  47. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/iterative_mean.py +0 -0
  48. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/meta_singleton.py +0 -0
  49. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/time_utils.py +0 -0
  50. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo/utils/timer.py +0 -0
  51. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/iapytoo.egg-info/dependency_links.txt +0 -0
  52. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/setup.cfg +0 -0
  53. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/tests/test_dataset.py +0 -0
  54. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/tests/test_metrics.py +0 -0
  55. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/tests/test_model.py +0 -0
  56. {iapytoo-0.0.2 → iapytoo-0.0.4.dev3}/tests/test_scaling.py +0 -0
@@ -0,0 +1,78 @@
1
+ # This workflow will install Python dependencies,
2
+ # run tests and lint with a single version of Python
3
+ # It will also publish on release branch or if tag v* is pushed
4
+
5
+ name: Python application
6
+
7
+ on:
8
+ push:
9
+ branches:
10
+ - main
11
+ - 'release/v*'
12
+ tags:
13
+ - 'v*'
14
+ pull_request:
15
+ branches: [ "main" ]
16
+
17
+ permissions:
18
+ contents: read
19
+
20
+ jobs:
21
+ build:
22
+
23
+ runs-on: ubuntu-latest
24
+
25
+ steps:
26
+ - uses: actions/checkout@v3
27
+ - name: Set up Python 3.10
28
+ uses: actions/setup-python@v3
29
+ with:
30
+ python-version: "3.10"
31
+ - name: Install dependencies
32
+ run: |
33
+ python -m pip install --upgrade pip
34
+ pip install flake8 pytest
35
+ if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
36
+ - name: Lint with flake8
37
+ run: |
38
+ # stop the build if there are Python syntax errors or undefined names
39
+ flake8 . --count --select=E9,F63,F7,F82 --show-source --statistics
40
+ # exit-zero treats all errors as warnings. The GitHub editor is 127 chars wide
41
+ flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
42
+ - name: set pythonpath
43
+ run: |
44
+ echo "----PYTHONPATH------"
45
+ PWD=$(pwd)
46
+ export PYTHONPATH=$PWD:$PYTHONPATH
47
+ echo "PYTHONPATH=$PYTHONPATH" >> $GITHUB_ENV
48
+ - name: Test with pytest
49
+ run: |
50
+ pytest
51
+
52
+ publish:
53
+ needs: build
54
+ if: startsWith(github.ref, 'refs/heads/release/v') || startsWith(github.ref, 'refs/tags/v')
55
+ runs-on: ubuntu-latest
56
+
57
+
58
+ steps:
59
+ - uses: actions/checkout@v3
60
+ with:
61
+ fetch-depth: 0 # Fetch all history for all tags and branches
62
+ - name: Set up Python
63
+ uses: actions/setup-python@v3
64
+ with:
65
+ python-version: '3.x'
66
+ - name: Install dependencies
67
+ run: |
68
+ python -m pip install --upgrade pip
69
+ pip install build setuptools setuptools-scm twine
70
+
71
+ - name: Build the package
72
+ run: python -m build
73
+
74
+ - name: Publish to PyPI
75
+ env:
76
+ TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
77
+ TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
78
+ run: twine upload dist/*
@@ -0,0 +1,168 @@
1
+ # Byte-compiled / optimized / DLL files
2
+ __pycache__/
3
+ */__pycache__/
4
+ *.py[cod]
5
+ *$py.class
6
+
7
+ # C extensions
8
+ *.so
9
+
10
+ # Distribution / packaging
11
+ .Python
12
+ build/
13
+ develop-eggs/
14
+ dist/
15
+ downloads/
16
+ eggs/
17
+ .eggs/
18
+ lib/
19
+ lib64/
20
+ parts/
21
+ sdist/
22
+ var/
23
+ wheels/
24
+ share/python-wheels/
25
+ *.egg-info/
26
+ .installed.cfg
27
+ *.egg
28
+ MANIFEST
29
+
30
+ # PyInstaller
31
+ # Usually these files are written by a python script from a template
32
+ # before PyInstaller builds the exe, so as to inject date/other infos into it.
33
+ *.manifest
34
+ *.spec
35
+
36
+ # Installer logs
37
+ pip-log.txt
38
+ pip-delete-this-directory.txt
39
+
40
+ # Unit test / coverage reports
41
+ htmlcov/
42
+ .tox/
43
+ .nox/
44
+ .coverage
45
+ .coverage.*
46
+ .cache
47
+ nosetests.xml
48
+ coverage.xml
49
+ *.cover
50
+ *.py,cover
51
+ .hypothesis/
52
+ .pytest_cache/
53
+ cover/
54
+
55
+ # Translations
56
+ *.mo
57
+ *.pot
58
+
59
+ # Django stuff:
60
+ *.log
61
+ local_settings.py
62
+ db.sqlite3
63
+ db.sqlite3-journal
64
+
65
+ # Flask stuff:
66
+ instance/
67
+ .webassets-cache
68
+
69
+ # Scrapy stuff:
70
+ .scrapy
71
+
72
+ # Sphinx documentation
73
+ docs/_build/
74
+
75
+ # PyBuilder
76
+ .pybuilder/
77
+ target/
78
+
79
+ # Jupyter Notebook
80
+ .ipynb_checkpoints
81
+
82
+ # IPython
83
+ profile_default/
84
+ ipython_config.py
85
+
86
+ # pyenv
87
+ # For a library or package, you might want to ignore these files since the code is
88
+ # intended to run in multiple environments; otherwise, check them in:
89
+ # .python-version
90
+
91
+ # pipenv
92
+ # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
93
+ # However, in case of collaboration, if having platform-specific dependencies or dependencies
94
+ # having no cross-platform support, pipenv may install dependencies that don't work, or not
95
+ # install all needed dependencies.
96
+ #Pipfile.lock
97
+
98
+ # poetry
99
+ # Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
100
+ # This is especially recommended for binary packages to ensure reproducibility, and is more
101
+ # commonly ignored for libraries.
102
+ # https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
103
+ #poetry.lock
104
+
105
+ # pdm
106
+ # Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
107
+ #pdm.lock
108
+ # pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
109
+ # in version control.
110
+ # https://pdm.fming.dev/#use-with-ide
111
+ .pdm.toml
112
+
113
+ # PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
114
+ __pypackages__/
115
+
116
+ # Celery stuff
117
+ celerybeat-schedule
118
+ celerybeat.pid
119
+
120
+ # SageMath parsed files
121
+ *.sage.py
122
+
123
+ # Environments
124
+ .env
125
+ .venv
126
+ env/
127
+ venv/
128
+ ENV/
129
+ env.bak/
130
+ venv.bak/
131
+
132
+ # Spyder project settings
133
+ .spyderproject
134
+ .spyproject
135
+
136
+ # Rope project settings
137
+ .ropeproject
138
+
139
+ # mkdocs documentation
140
+ /site
141
+
142
+ # mypy
143
+ .mypy_cache/
144
+ .dmypy.json
145
+ dmypy.json
146
+
147
+ # Pyre type checker
148
+ .pyre/
149
+
150
+ # pytype static type analyzer
151
+ .pytype/
152
+
153
+ # Cython debug symbols
154
+ cython_debug/
155
+
156
+ # PyCharm
157
+ # JetBrains specific template is maintained in a separate JetBrains.gitignore that can
158
+ # be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
159
+ # and can be added to the global gitignore or merged into this file. For a more nuclear
160
+ # option (not recommended) you can uncomment the following to ignore the entire idea folder.
161
+ .idea/
162
+
163
+ # VS Code
164
+ .vscode
165
+
166
+
167
+ mlruns/
168
+ data/
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.2
2
2
  Name: iapytoo
3
- Version: 0.0.2
3
+ Version: 0.0.4.dev3
4
4
  Summary: Un toolkit pour vite coder un modèle pytorch d'IA utilisant MLflow
5
5
  Author-email: Jean-François Lecomte <zwergon@free.fr>
6
6
  License: Apache-2.0
@@ -0,0 +1,13 @@
1
+ FROM nvidia/cuda:12.2.0-runtime-ubuntu22.04
2
+
3
+ RUN apt-get update && apt-get install -y \
4
+ curl \
5
+ git \
6
+ python3 \
7
+ python3-pip
8
+
9
+ RUN pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu121
10
+
11
+ # Install pip dependencies
12
+ COPY requirements.txt .
13
+ RUN pip install --no-cache-dir -r requirements.txt
@@ -0,0 +1,7 @@
1
+
2
+ # Création de l'image Docker.
3
+
4
+ Dans le répertoire `.` (racine) de iapytoo
5
+ ```
6
+ docker build -t iapytoo -f dockers/Dockerfile .
7
+ ```
@@ -0,0 +1,9 @@
1
+ #!/bin/bash
2
+
3
+ docker run --rm --gpus all \
4
+ -v "$(pwd)":/app \
5
+ -w /app \
6
+ -e PYTHONPATH=/app \
7
+ iapytoo:latest \
8
+ python3 examples/mnist.py -c examples/config_mnist.json data
9
+
@@ -0,0 +1,19 @@
1
+ {
2
+ "project": "mnist",
3
+ "run": "mnist",
4
+ "tracking_uri": null,
5
+ "tqdm": true,
6
+ "normalization": true,
7
+ "n_steps_by_batch": 10,
8
+ "ratio_train_test": 0.9,
9
+ "num_workers": 2,
10
+ "loss": "nll",
11
+ "cuda": true,
12
+ "model": "mnist",
13
+ "optimizer": "adam",
14
+ "scheduler": "step",
15
+ "gamma": 0.7,
16
+ "seed": 1,
17
+ "batch_size": 64,
18
+ "epochs": 11
19
+ }
@@ -0,0 +1,25 @@
1
+ {
2
+ "project": "wgan",
3
+ "run": "wgan",
4
+ "data_dir": "./data",
5
+ "tracking_uri": null,
6
+ "tqdm": true,
7
+ "normalization": true,
8
+ "n_steps_by_batch": 10,
9
+ "ratio_train_test": 0.9,
10
+ "num_workers": 2,
11
+ "cuda": true,
12
+ "generator": "generator",
13
+ "discriminator": "critic",
14
+ "n_critic": 1,
15
+ "optimizer": "adam",
16
+ "scheduler": "step",
17
+ "gamma": 0.7,
18
+ "betas": [0.5, 0.99],
19
+ "lambda_gp": 10,
20
+ "noise_dim": 100,
21
+ "seed": 1,
22
+ "batch_size": 256,
23
+ "epochs": 11,
24
+ "learning_rate": 0.0002
25
+ }
@@ -0,0 +1,106 @@
1
+ import torch
2
+ import torch.nn as nn
3
+ import torch.nn.functional as F
4
+ from torchvision import datasets, transforms
5
+ from torch.optim.lr_scheduler import LambdaLR
6
+
7
+
8
+ from iapytoo.dataset.scaling import Scaling
9
+ from iapytoo.predictions.predictors import MaxPredictor
10
+ from iapytoo.predictions.plotters import ConfusionPlotter
11
+ from iapytoo.train.factories import Model, ModelFactory, SchedulerFactory, Scheduler
12
+ from iapytoo.utils.config import Config
13
+ from iapytoo.train.training import Training
14
+
15
+
16
+ import matplotlib.pyplot as plt
17
+
18
+
19
+ class MnistModel(Model):
20
+ def __init__(self, loader, config):
21
+ super(MnistModel, self).__init__(loader, config)
22
+ self.conv1 = nn.Conv2d(1, 32, 3, 1)
23
+ self.conv2 = nn.Conv2d(32, 64, 3, 1)
24
+ self.dropout1 = nn.Dropout(0.25)
25
+ self.dropout2 = nn.Dropout(0.5)
26
+ self.fc1 = nn.Linear(9216, 128)
27
+ self.fc2 = nn.Linear(128, 10)
28
+
29
+ def forward(self, x):
30
+ x = self.conv1(x)
31
+ x = F.relu(x)
32
+ x = self.conv2(x)
33
+ x = F.relu(x)
34
+ x = F.max_pool2d(x, 2)
35
+ x = self.dropout1(x)
36
+ x = torch.flatten(x, 1)
37
+ x = self.fc1(x)
38
+ x = F.relu(x)
39
+ x = self.dropout2(x)
40
+ x = self.fc2(x)
41
+ output = F.log_softmax(x, dim=1)
42
+ return output
43
+
44
+ def predict(self, model_output):
45
+ return torch.argmax(model_output, dim=1)
46
+
47
+
48
+ class MnistScheduler(Scheduler):
49
+ def __init__(self, optimizer, config) -> None:
50
+ super().__init__(optimizer, config)
51
+
52
+ def lr_lambda(epoch):
53
+ # LR to be 0.1 * (1/1+0.01*epoch)
54
+ return 0.995**epoch
55
+
56
+ self.lr_scheduler = LambdaLR(optimizer, lr_lambda=lr_lambda)
57
+
58
+
59
+ class MnistTraining(Training):
60
+ def __init__(self, config: Config) -> None:
61
+ super().__init__(config, predictor=MaxPredictor(), metric_names=[])
62
+ self.predictions.add_plotter(ConfusionPlotter())
63
+
64
+
65
+ if __name__ == "__main__":
66
+ import os
67
+
68
+ from iapytoo.utils.arguments import parse_args
69
+
70
+ ModelFactory().register_model("mnist", MnistModel)
71
+ SchedulerFactory().register_scheduler("mnist", MnistScheduler)
72
+
73
+ args = parse_args()
74
+
75
+ if args.run_id is not None:
76
+ config = Config.create_from_run_id(args.run_id, args.tracking_uri)
77
+ config.epochs = args.epochs
78
+ else:
79
+ # INPUT Parameters
80
+ config = Config.create_from_args(args)
81
+
82
+ Training.seed(config)
83
+
84
+ config = Config.create_from_args(args)
85
+
86
+ training = MnistTraining(config)
87
+
88
+ transform = transforms.Compose(
89
+ [transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]
90
+ )
91
+
92
+ dataset1 = datasets.MNIST(
93
+ args.dataset, train=True, download=True, transform=transform
94
+ )
95
+ dataset2 = datasets.MNIST(args.dataset, train=False, transform=transform)
96
+
97
+ train_loader = torch.utils.data.DataLoader(
98
+ dataset1, batch_size=config.batch_size, num_workers=config.num_workers
99
+ )
100
+ test_loader = torch.utils.data.DataLoader(
101
+ dataset2, batch_size=config.batch_size, num_workers=config.num_workers
102
+ )
103
+
104
+ training.fit(
105
+ train_loader=train_loader, valid_loader=test_loader, run_id=args.run_id
106
+ )
@@ -0,0 +1,200 @@
1
+ import os
2
+ import torch
3
+ import torch.nn as nn
4
+ import torch.optim as optim
5
+ import numpy as np
6
+ from torch.autograd import grad
7
+
8
+ from iapytoo.train.wgan import WGAN
9
+ from iapytoo.utils.config import Config
10
+
11
+
12
+ from iapytoo.train.factories import Model, ModelFactory, OptimizerFactory
13
+ from iapytoo.predictions.plotters import PredictionPlotter, Fake2DPlotter
14
+
15
+ import torch
16
+ import torch.nn as nn
17
+ from torch.utils.data import Dataset
18
+ import torchvision
19
+ import numpy as np
20
+ from torchvision.datasets import MNIST
21
+ import torchvision.transforms as transforms
22
+ import matplotlib.pyplot as plt
23
+ from iapytoo.train.factories import WeightInitiator
24
+
25
+
26
+ class MNISTInitiator(WeightInitiator):
27
+ def __call__(self, m):
28
+ if isinstance(m, nn.Conv2d) or isinstance(m, nn.ConvTranspose2d):
29
+ torch.nn.init.normal_(m.weight, 0.0, 0.02)
30
+ if isinstance(m, nn.BatchNorm2d):
31
+ torch.nn.init.normal_(m.weight, 0.0, 0.02)
32
+ torch.nn.init.constant_(m.bias, 0)
33
+
34
+
35
+ class Generator(Model):
36
+ """See https://github.com/Zeleni9/pytorch-wgan"""
37
+
38
+ @staticmethod
39
+ def get_noise(n_samples, z_dim, device="cpu"):
40
+ """
41
+ Function for creating noise vectors: Given the dimensions (n_samples, z_dim)
42
+ creates a tensor of that shape filled with random numbers from the normal distribution.
43
+ Parameters:
44
+ n_samples: the number of samples to generate, a scalar
45
+ z_dim: the dimension of the noise vector, a scalar
46
+ device: the device type
47
+ """
48
+ return torch.randn(n_samples, z_dim, device=device)
49
+
50
+ def __init__(self, loader, config):
51
+ super(Generator, self).__init__(loader, config)
52
+ self.z_dim = config["noise_dim"]
53
+ # Filters [1024, 512, 256]
54
+ # Input_dim = 100
55
+ # Output_dim = 1 (number of channels)
56
+ self.main_module = nn.Sequential(
57
+ # Z latent vector 100
58
+ nn.ConvTranspose2d(
59
+ in_channels=self.z_dim,
60
+ out_channels=1024,
61
+ kernel_size=4,
62
+ stride=1,
63
+ padding=0,
64
+ ),
65
+ nn.BatchNorm2d(num_features=1024),
66
+ nn.ReLU(True),
67
+ # State (1024x4x4)
68
+ nn.ConvTranspose2d(
69
+ in_channels=1024, out_channels=512, kernel_size=4, stride=2, padding=1
70
+ ),
71
+ nn.BatchNorm2d(num_features=512),
72
+ nn.ReLU(True),
73
+ # State (512x8x8)
74
+ nn.ConvTranspose2d(
75
+ in_channels=512, out_channels=256, kernel_size=4, stride=2, padding=1
76
+ ),
77
+ nn.BatchNorm2d(num_features=256),
78
+ nn.ReLU(True),
79
+ # State (256x16x16)
80
+ nn.ConvTranspose2d(
81
+ in_channels=256, out_channels=1, kernel_size=4, stride=2, padding=1
82
+ ),
83
+ )
84
+ # output of main module --> Image (Cx32x32)
85
+
86
+ self.output = nn.Tanh()
87
+
88
+ def unsqueeze_noise(self, noise):
89
+ """
90
+ Function for completing a forward pass of the generator: Given a noise tensor,
91
+ returns a copy of that noise with width and height = 1 and channels = z_dim.
92
+ Parameters:
93
+ noise: a noise tensor with dimensions (n_samples, z_dim)
94
+ """
95
+ return noise.view(noise.shape[0], self.z_dim, 1, 1)
96
+
97
+ def forward(self, noise):
98
+ """
99
+ Function for completing a forward pass of the generator: Given a noise tensor,
100
+ returns generated images.
101
+ Parameters:
102
+ noise: a noise tensor with dimensions (n_samples, z_dim)
103
+ """
104
+ x = self.unsqueeze_noise(noise)
105
+ x = self.main_module(x)
106
+ return self.output(x)
107
+
108
+
109
+ class Discriminator(Model):
110
+ def __init__(self, loader, config):
111
+ super(Discriminator, self).__init__(loader, config)
112
+ # Filters [256, 512, 1024]
113
+ # Input_dim = 1 (Cx64x64)
114
+ # Output_dim = 1
115
+ self.main_module = nn.Sequential(
116
+ # Omitting batch normalization in critic because our new penalized training objective (WGAN with gradient penalty) is no longer valid
117
+ # in this setting, since we penalize the norm of the critic's gradient with respect to each input independently and not the enitre batch.
118
+ # There is not good & fast implementation of layer normalization --> using per instance normalization nn.InstanceNorm2d()
119
+ # Image (Cx32x32)
120
+ nn.Conv2d(
121
+ in_channels=1, out_channels=256, kernel_size=4, stride=2, padding=1
122
+ ),
123
+ nn.InstanceNorm2d(256, affine=True),
124
+ nn.LeakyReLU(0.2, inplace=True),
125
+ # State (256x16x16)
126
+ nn.Conv2d(
127
+ in_channels=256, out_channels=512, kernel_size=4, stride=2, padding=1
128
+ ),
129
+ nn.InstanceNorm2d(512, affine=True),
130
+ nn.LeakyReLU(0.2, inplace=True),
131
+ # State (512x8x8)
132
+ nn.Conv2d(
133
+ in_channels=512, out_channels=1024, kernel_size=4, stride=2, padding=1
134
+ ),
135
+ nn.InstanceNorm2d(1024, affine=True),
136
+ nn.LeakyReLU(0.2, inplace=True),
137
+ )
138
+ # output of main module --> State (1024x4x4)
139
+
140
+ self.output = nn.Sequential(
141
+ # The output of D is no longer a probability, we do not apply sigmoid at the output of D.
142
+ nn.Conv2d(
143
+ in_channels=1024, out_channels=1, kernel_size=4, stride=1, padding=0
144
+ )
145
+ )
146
+
147
+ def forward(self, x):
148
+ x = self.main_module(x)
149
+ return self.output(x)
150
+
151
+ def feature_extraction(self, x):
152
+ # Use discriminator for feature extraction then flatten to vector of 16384
153
+ x = self.main_module(x)
154
+ return x.view(-1, 1024 * 4 * 4)
155
+
156
+
157
+ class LatentDataset(Dataset):
158
+ def __init__(self, noise_dim, size) -> None:
159
+ super().__init__()
160
+ self.noise = torch.randn(size, noise_dim)
161
+
162
+ def __len__(self):
163
+ return self.noise.shape[0]
164
+
165
+ def __getitem__(self, idx):
166
+ return self.noise[idx, :]
167
+
168
+
169
+ if __name__ == "__main__":
170
+ config = Config(os.path.join(os.path.dirname(__file__), "config_wgan.json"))
171
+
172
+ transform = transforms.Compose(
173
+ [
174
+ transforms.Resize(32),
175
+ transforms.ToTensor(),
176
+ transforms.Normalize((0.5,), (0.5,)),
177
+ ]
178
+ )
179
+
180
+ # load training data
181
+ trainset = MNIST(config.data_dir, download=True, train=True, transform=transform)
182
+
183
+ trainloader = torch.utils.data.DataLoader(
184
+ trainset, batch_size=config.batch_size, shuffle=True, drop_last=True
185
+ )
186
+
187
+ latentset = LatentDataset(config.noise_dim, size=16)
188
+
189
+ valid_loader = torch.utils.data.DataLoader(latentset, batch_size=1, shuffle=False)
190
+
191
+ item = next(iter(valid_loader))
192
+ print(item.shape)
193
+
194
+ model_factory = ModelFactory()
195
+ model_factory.register_model("generator", Generator)
196
+ model_factory.register_model("critic", Discriminator)
197
+
198
+ wgan = WGAN(config)
199
+ wgan.predictions.add_plotter(Fake2DPlotter())
200
+ wgan.fit(train_loader=trainloader, valid_loader=valid_loader, run_id=None)
@@ -0,0 +1,16 @@
1
+ # file generated by setuptools_scm
2
+ # don't change, don't track in version control
3
+ TYPE_CHECKING = False
4
+ if TYPE_CHECKING:
5
+ from typing import Tuple, Union
6
+ VERSION_TUPLE = Tuple[Union[int, str], ...]
7
+ else:
8
+ VERSION_TUPLE = object
9
+
10
+ version: str
11
+ __version__: str
12
+ __version_tuple__: VERSION_TUPLE
13
+ version_tuple: VERSION_TUPLE
14
+
15
+ __version__ = version = '0.0.4.dev3'
16
+ __version_tuple__ = version_tuple = (0, 0, 4, 'dev3')
@@ -0,0 +1,17 @@
1
+ {
2
+ "project": "iapytoo",
3
+ "run": "run",
4
+ "tracking_uri": null,
5
+ "tqdm": true,
6
+ "normalization": true,
7
+ "n_steps_by_batch": 10,
8
+ "ratio_train_test": 0.9,
9
+ "num_workers": 2,
10
+ "model": "CNN",
11
+ "kernel_size": 3,
12
+ "dropout": 0,
13
+ "loss": "mse",
14
+ "optimizer": "adam",
15
+ "cuda": true,
16
+ "seed":true
17
+ }
@@ -0,0 +1,17 @@
1
+ {
2
+ "project": "iapytoo_test",
3
+ "run": "run",
4
+ "tracking_uri": null,
5
+ "tqdm": true,
6
+ "normalization": true,
7
+ "n_steps_by_batch": 10,
8
+ "ratio_train_test": 0.9,
9
+ "num_workers": 2,
10
+ "model": "CNN",
11
+ "kernel_size": 3,
12
+ "dropout": 0,
13
+ "loss": "mse",
14
+ "optimizer": "adam",
15
+ "cuda": true,
16
+ "seed":true
17
+ }
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.2
2
2
  Name: iapytoo
3
- Version: 0.0.2
3
+ Version: 0.0.4.dev3
4
4
  Summary: Un toolkit pour vite coder un modèle pytorch d'IA utilisant MLflow
5
5
  Author-email: Jean-François Lecomte <zwergon@free.fr>
6
6
  License: Apache-2.0
@@ -1,8 +1,19 @@
1
+ .gitignore
1
2
  LICENSE
2
3
  MANIFEST.in
3
4
  README.md
4
5
  pyproject.toml
6
+ requirements.txt
7
+ .github/workflows/python-app.yml
8
+ dockers/Dockerfile
9
+ dockers/Readme.md
10
+ dockers/run-python.sh
11
+ examples/config_mnist.json
12
+ examples/config_wgan.json
13
+ examples/mnist.py
14
+ examples/wgan_mnist.py
5
15
  iapytoo/__init__.py
16
+ iapytoo/_version.py
6
17
  iapytoo.egg-info/PKG-INFO
7
18
  iapytoo.egg-info/SOURCES.txt
8
19
  iapytoo.egg-info/dependency_links.txt
@@ -34,6 +45,8 @@ iapytoo/utils/iterative_mean.py
34
45
  iapytoo/utils/meta_singleton.py
35
46
  iapytoo/utils/time_utils.py
36
47
  iapytoo/utils/timer.py
48
+ iapytoo/utils/cf/config.json
49
+ iapytoo/utils/cf/config_test.json
37
50
  tests/test_dataset.py
38
51
  tests/test_metrics.py
39
52
  tests/test_model.py
@@ -0,0 +1,3 @@
1
+ dist
2
+ dockers
3
+ iapytoo
@@ -1,10 +1,10 @@
1
1
  [build-system]
2
- requires = ["setuptools>=61.0"]
2
+ requires = ["setuptools>=61.0", "setuptools_scm>=8"]
3
3
  build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "iapytoo"
7
- version = "0.0.2"
7
+ dynamic = ["version"]
8
8
  description = "Un toolkit pour vite coder un modèle pytorch d'IA utilisant MLflow"
9
9
 
10
10
  authors = [
@@ -13,20 +13,16 @@ authors = [
13
13
 
14
14
  license = { text = "Apache-2.0" }
15
15
 
16
-
17
16
  readme = "README.md"
18
17
  requires-python = ">=3.7"
19
18
 
20
- [tool.setuptools]
21
- packages = ["iapytoo",
22
- "iapytoo.dataset",
23
- "iapytoo.metrics",
24
- "iapytoo.utils",
25
- "iapytoo.predictions",
26
- "iapytoo.train"
27
- ]
19
+ [tool.setuptools_scm]
20
+ version_file = "iapytoo/_version.py"
21
+ local_scheme = "no-local-version"
28
22
 
29
- license-files = ["LICENSE"]
23
+ [tool.setuptools.packages.find]
24
+ where = ["."]
25
+ exclude = ["examples", "tests"]
30
26
 
31
27
  [tool.setuptools.dynamic]
32
28
  dependencies = {file = ["requirements.txt"]}
@@ -0,0 +1,6 @@
1
+ tqdm
2
+ numpy
3
+ torch==2.4.1
4
+ torchvision==0.19.1
5
+ mlflow
6
+ seaborn
@@ -1 +0,0 @@
1
- iapytoo
File without changes
File without changes
File without changes
File without changes