zChainer
scikit-learn like interface and stacked autoencoder for chainer
Requirements
numpy
scikit-learn
chainer >= 1.5
Installation
pip install zChainer
Usage
Autoencoder
import numpy as np
import chainer.functions as F
import chainer.links as L
from chainer import ChainList, optimizers
from zChainer import NNAutoEncoder, utility
data = (..).astype(np.float32)
encoder = ChainList(
L.Linear(784, 200),
L.Linear(200, 100))
decoder =ChainList(
L.Linear(200, 784),
L.Linear(100, 200))
# You can set your own forward function. Default is as below.
#def forward(self, x):
# h = F.dropout(F.relu(self.model[0](x)))
# return F.dropout(F.relu(self.model[1](h)))
#
#NNAutoEncoder.forward = forward
ae = NNAutoEncoder(encoder, decoder, optimizers.Adam(), epoch=100, batch_size=100,
log_path="./ae_log_"+utility.now()+".csv", export_path="./ae_"+utility.now()+".model")
ae.fit(data)Training and Testing
import numpy as np
import chainer.functions as F
import chainer.links as L
from chainer import ChainList, optimizers
from zChainer import NNManager, utility
import pickle
X_train = (..).astype(np.float32)
y_train = (..).astype(np.int32)
X_test = (..).astype(np.float32)
y_test = (..).astype(np.int32)
# Create a new network
model = ChainList(L.Linear(784, 200), L.Linear(200, 100), L.Linear(100, 10))
# or load a serialized model
#f = open("./ae_2015-12-01_11-26-45.model")
#model = pickle.load(f)
#f.close()
#model.add_link(L.Linear(100,10))
def forward(self, x):
h = F.relu(self.model[0](x))
h = F.relu(self.model[1](h))
return F.relu(self.model[2](h))
def output(self, y):
y_trimed = y.data.argmax(axis=1)
return np.array(y_trimed, dtype=np.int32)
NNManager.forward = forward
NNManager.output = output
nn = NNManager(model, optimizers.Adam(), F.softmax_cross_entropy,
epoch=100, batch_size=100,
log_path="./training_log_"+utility.now()+".csv")
nn.fit(X_train, y_train, is_classification=True)
nn.predict(X_test, y_test)1 maintainer
Metadata
1 maintainer
Release files for zChainer 0.3.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| zChainer-0.3.2.tar.gz | 3.7 kB | Without provenance | Details |
Release files / zChainer-0.3.2.tar.gz
| Download URL | zChainer-0.3.2.tar.gz |
|---|---|
| Size | 3.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
ceebcacb13f5bf000452e0fdb099bd6c706af43ff0c7df2227109ecb32495fa1
|
|
BLAKE2b-256 checksum How to use checksums |
a38bc5098e686351297db0d79a834f3d82a2365e6a05beb709068ab07871f9fc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
Security for zChainer 0.3.2
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
