Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/tailhq/DynaML
Scala Library/REPL for Machine Learning Research
https://github.com/tailhq/DynaML
classification committee-models gaussian-processes kernel-methods machine-learning machine-learning-algorithms machine-learning-api regression repl scala scala-library tensorflow
Last synced: 7 days ago
JSON representation
Scala Library/REPL for Machine Learning Research
- Host: GitHub
- URL: https://github.com/tailhq/DynaML
- Owner: tailhq
- License: apache-2.0
- Created: 2015-02-16T15:22:30.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2023-04-21T20:44:11.000Z (over 1 year ago)
- Last Synced: 2024-05-29T02:05:51.276Z (5 months ago)
- Topics: classification, committee-models, gaussian-processes, kernel-methods, machine-learning, machine-learning-algorithms, machine-learning-api, regression, repl, scala, scala-library, tensorflow
- Language: Scala
- Homepage: http://tailhq.github.io/DynaML/
- Size: 215 MB
- Stars: 198
- Watchers: 19
- Forks: 51
- Open Issues: 25
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
Awesome Lists containing this project
- awesome - DynaML - Scala Library/REPL for Machine Learning Research (Scala)
README
# DynaML: ML + JVM + Scala
[](https://gitter.im/DynaML/Lobby?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)  [](https://jitpack.io/#transcendent-ai-labs/DynaML)
[](https://coveralls.io/github/transcendent-ai-labs/DynaML)
------------------
DynaML is a Scala & JVM Machine Learning toolbox for research, education & industry.
------------------
## Motivation
- __Interactive.__ Don't want to create Maven/sbt project skeletons
every time you want to try out ideas? Create and execute [scala worksheets](scripts/randomvariables.sc)
in the DynaML shell. DynaML comes packaged with a customized version of the [Ammonite](http://ammonite.io) REPL,
with *auto-complete*, file operations and scripting capabilities.
- __End to End.__ Create complex pre-processing pipelines with the [data pipes](https://transcendent-ai-labs.github.io/DynaML/pipes/pipes/) API,
train models ([deep nets](scripts/cifar.sc), [gaussian processes](https://transcendent-ai-labs.github.io/DynaML/core/core_gp/),
[linear models](https://transcendent-ai-labs.github.io/DynaML/core/core_glm/) and more),
optimize over [hyper-parameters](https://transcendent-ai-labs.github.io/DynaML/core/core_opt_global/),
[evaluate](https://transcendent-ai-labs.github.io/DynaML/core/core_model_evaluation/) model predictions and
[visualise](https://transcendent-ai-labs.github.io/DynaML/core/core_graphics/) results.
- __Enterprise Friendly.__ Take advantage of the JVM and Scala ecosystem, use Apache [Spark](https://spark.apache.org)
to write scalable data analysis jobs, [Tensorflow](http://tensorflow.org) for deep learning, all in the same toolbox.
------------------
## Getting Started
### Platform Compatibility
Currently, only *nix and OSX platforms are supported.
DynaML is compatible with Scala `2.12`
### Installation
Easiest way to install DynaML is cloning & compiling from the [github](/) repository. Please take a look at
the [installation](https://transcendent-ai-labs.github.io/DynaML/installation/installation/) instructions in the
[user guide](https://transcendent-ai-labs.github.io/DynaML/), to make sure that you have the pre-requisites
and to configure your installation.
------------------
## CIFAR in under 200 lines
Below is a sample [script](scripts/cifar.sc) where we train a neural network of stacked
[Inception](https://arxiv.org/pdf/1409.4842.pdf) cells on the [CIFAR-10](https://en.wikipedia.org/wiki/CIFAR-10)
image classification task.
```scala
{
import _root_.ammonite.ops._
import _root_.io.github.tailhq.dynaml.pipes.DataPipe
import _root_.io.github.tailhq.dynaml.tensorflow.{
dtflearn,
dtfutils,
dtfdata,
dtfpipe
}
import _root_.org.platanios.tensorflow.api._
import _root_.org.platanios.tensorflow.api.learn.layers.Activation
import _root_.org.platanios.tensorflow.data.image.CIFARLoader
import _root_.java.nio.file.Paths
val tempdir = home / "tmp"
val dataSet =
CIFARLoader.load(Paths.get(tempdir.toString()), CIFARLoader.CIFAR_10)
val dtf_cifar_data = dtfdata.tf_dataset(
dtfdata.supervised_dataset(
dataSet.trainImages.unstack(axis = 0),
dataSet.trainLabels.castTo[Long].unstack(axis = -1)
),
dtfdata.supervised_dataset(
dataSet.testImages.unstack(axis = 0),
dataSet.testLabels.castTo[Long].unstack(axis = -1)
)
)
println("Building the model.")
val relu_act =
DataPipe[String, Activation[Float]]((x: String) => tf.learn.ReLU[Float](x))
val architecture =
tf.learn.Cast[UByte, Float]("Input/Cast") >>
dtflearn.inception_unit[Float](channels = 3, Seq.fill(4)(10), relu_act)(
layer_index = 1
) >>
dtflearn.inception_unit[Float](channels = 40, Seq.fill(4)(5), relu_act)(
layer_index = 2
) >>
tf.learn.Flatten[Float]("Layer_3/Flatten") >>
dtflearn.feedforward[Float](256)(id = 4) >>
tf.learn.ReLU[Float]("Layer_4/ReLU", 0.1f) >>
dtflearn.feedforward[Float](10)(id = 5)
val loss = tf.learn.SparseSoftmaxCrossEntropy[Float, Long, Float](
"Loss/CrossEntropy"
) >>
tf.learn.Mean("Loss/Mean") >>
tf.learn.ScalarSummary("Loss/Summary", "Loss")
val optimizer = tf.train.Adam(0.1f)
val cifar_model =
dtflearn.model[
Output[UByte], Output[Long], Output[Float], Float,
Tensor[UByte], UINT8, Shape,
Tensor[Long], INT64, Shape,
Tensor[Float], FLOAT32, Shape](
architecture,
(UINT8, dataSet.trainImages.shape(1 ::)),
(INT64, Shape()),
loss
)
val data_ops = dtflearn.model.data_ops[(Output[UByte], Output[Long])](
shuffleBuffer = 5000,
batchSize = 128,
prefetchSize = 10
)
val train_config = dtflearn.model.trainConfig(
tempdir / "cifar_summaries",
data_ops,
optimizer,
dtflearn.rel_loss_change_stop(0.05, 500),
Some(
dtflearn.model._train_hooks(
tempdir / "cifar_summaries",
stepRateFreq = 100,
summarySaveFreq = 100,
checkPointFreq = 100
)
)
)
val pattern_to_tensor =
DataPipe[Seq[(Tensor[UByte], Tensor[Long])], (Tensor[UByte], Tensor[Long])](
ds => {
val (xs, ys) = ds.unzip
(
dtfpipe.EagerStack[UByte](axis = 0).run(xs),
dtfpipe.EagerStack[Long](axis = 0).run(ys)
)
}
)
val data_handle_ops = dtflearn.model.tf_data_handle_ops[
(Tensor[UByte], Tensor[Long]),
(Tensor[UByte], Tensor[Long]),
Tensor[Float],
(Output[UByte], Output[Long])
](
bufferSize = 500,
patternToTensor = Some(pattern_to_tensor),
concatOpO = Some(dtfpipe.EagerConcatenate[Float]())
)
val data_handle_ops_infer =
dtflearn.model.tf_data_handle_ops[Tensor[UByte], Tensor[UByte], Tensor[
Float
], Output[UByte]](
bufferSize = 1000,
patternToTensor = Some(dtfpipe.EagerStack[UByte](axis = 0)),
concatOpO = Some(dtfpipe.EagerConcatenate[Float]())
)
cifar_model.train(
dtf_cifar_data.training_dataset,
train_config,
data_handle_ops
)
def accuracy(predictions: Tensor[Long], labels: Tensor[Long]): Float =
tfi
.equal(predictions.argmax[Long](1), labels)
.castTo[Float]
.mean()
.scalar
.asInstanceOf[Float]
val (trainingPreds, testPreds): (Tensor[Float], Tensor[Float]) = (
cifar_model
.infer_batch(
dtf_cifar_data.training_dataset.map(p => p._1),
data_handle_ops_infer
)
.left
.get,
cifar_model
.infer_batch(
dtf_cifar_data.test_dataset.map(p => p._1),
data_handle_ops_infer
)
.left
.get
)
val (trainAccuracy, testAccuracy) = (
accuracy(trainingPreds.castTo[Long], dataSet.trainLabels.castTo[Long]),
accuracy(testPreds.castTo[Long], dataSet.testLabels.castTo[Long])
)
print("Train accuracy = ")
pprint.pprintln(trainAccuracy)
print("Test accuracy = ")
pprint.pprintln(testAccuracy)
}
```
------------------
## Support & Community
- [User guide](https://transcendent-ai-labs.github.io/DynaML/)
- [Gitter](https://gitter.im/DynaML/Lobby?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
- [Contributing](https://github.com/transcendent-ai-labs/DynaML/blob/master/CONTRIBUTING.md)
- [Code of Conduct](https://github.com/transcendent-ai-labs/DynaML/blob/master/CODE_OF_CONDUCT.md)