def setUpClass(cls):
cls.model_dir = tempfile.TemporaryDirectory()
cls.model = PyTorchNeuralNetwork(
classifications=["rock", "paper", "scissors"],
features=Features(Feature("image", int, 300 * 300)),
predict=Feature("label", int, 1),
directory=cls.model_dir.name,
network=RockPaperScissorsModel,
epochs=1,
batch_size=32,
imageSize=150,
validation_split=0.2,
loss=Loss,
optimizer="Adam",
enableGPU=True,
)
def setUpClass(cls):
cls.model_dir = tempfile.TemporaryDirectory()
cls.model = PyTorchNeuralNetwork(
classifications=["rock", "paper", "scissors"],
features=Features(Feature("image", int, 300 * 300)),
predict=Feature("label", int, 1),
directory=cls.model_dir.name,
network=RockPaperScissorsModel,
epochs=1,
batch_size=32,
imageSize=150,
validation_split=0.2,
loss=Loss,
optimizer="Adam",
enableGPU=True,
)
cls.traindir = asyncio.run(
cached_download_unpack_archive(
"https://storage.googleapis.com/laurencemoroney-blog.appspot.com/rps.zip",
"rps.zip",
"traindir",
"c6a9119b0c6a0907b782bd99e04ce09a0924c0895df6a26bc6fb06baca4526f55e51f7156cceb4791cc65632d66085e8",
)
)
cls.testdir = asyncio.run(
cached_download_unpack_archive(
"https://storage.googleapis.com/laurencemoroney-blog.appspot.com/rps-test-set.zip",
"rps-test-set.zip",
"testdir",
"fc45a0ebe58b9aafc3cd5a60020fa042d3a19c26b0f820aee630b9602c8f53dd52fd40f35d44432dd031dea8f30a5f66",
)
)
cls.predictdir = asyncio.run(
cached_download_unpack_archive(
"https://storage.googleapis.com/laurencemoroney-blog.appspot.com/rps-validation.zip",
"rps-validation.zip",
"predictdir",
"375457bb95771ffeace2beedab877292d232f31e76502618d25e0d92a3e029d386429f52c771b05ae1c7229d2f5ecc29",
)
)
x = self.linear(x.view(-1, 16 * 9 * 9))
return x
RockPaperScissorsModel = ConvNet()
Loss = CrossEntropyLossFunction()
# Define the dffml model config
model = PyTorchNeuralNetwork(
classifications=["rock", "paper", "scissors"],
features=Features(Feature("image", int, 300 * 300)),
predict=Feature("label", int, 1),
directory="rps_model",
network=RockPaperScissorsModel,
epochs=10,
batch_size=32,
imageSize=150,
validation_split=0.2,
loss=Loss,
optimizer="Adam",
enableGPU=True,
patience=2,
)
# Define source for training image dataset
train_source = DirectorySource(
foldername="rps",
feature="image",
labels=["rock", "paper", "scissors"],
)