Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[SPARKNLP-941] Adding OnnxSession to broadcast onnx options #14078

Merged
merged 4 commits into from
Dec 7, 2023
Merged

[SPARKNLP-941] Adding OnnxSession to broadcast onnx options #14078

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
1db94d2
[SPARKNLP-941] Adding OnnxSession to broadcast Onnx session option in…
danilojsl Nov 27, 2023
3fc8eeb
[SPARKNLP-941] Adding tensorflow installation to notebooks
danilojsl Nov 29, 2023
ee02641
Update HuggingFace_ONNX_in_Spark_NLP_Whisper.ipynb
maziyarpanahi Dec 7, 2023
2ad9fe3
Merge branch 'release/520-release-candidate' into bug/SPARKNLP-941-Sp…
danilojsl Dec 7, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion ...s/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_AlbertForQuestionAnswering.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@
}
],
"source": [
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum sentencepiece"
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum sentencepiece tensorflow"
]
},
{
Expand Down
18 changes: 14 additions & 4 deletions examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_BERT.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,7 @@
}
],
"source": [
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum"
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum tensorflow"
]
},
{
Expand Down Expand Up @@ -497,11 +497,21 @@
"provenance": []
},
"kernelspec": {
"display_name": "Python 3",
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python"
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.12"
},
"widgets": {
"application/vnd.jupyter.widget-state+json": {
Expand Down Expand Up @@ -2219,5 +2229,5 @@
}
},
"nbformat": 4,
"nbformat_minor": 0
"nbformat_minor": 1
}
2 changes: 1 addition & 1 deletion ...les/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_BertForQuestionAnswering.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
"source": [
"![JohnSnowLabs](https://sparknlp.org/assets/images/logo.png)\n",
"\n",
"[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/JohnSnowLabs/spark-nlp/blob/master/examples/python/transformers/onnx/HuggingFace%20ONNX%20in%20Spark%20NLP%20-%20AlbertForQuestionAnswering.ipynb)"
"[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/JohnSnowLabs/spark-nlp/blob/master/examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_BertForQuestionAnswering.ipynb)"
]
},
{
Expand Down
18 changes: 14 additions & 4 deletions examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_DistilBERT.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,7 @@
}
],
"source": [
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum"
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum tensorflow"
]
},
{
Expand Down Expand Up @@ -503,11 +503,21 @@
"provenance": []
},
"kernelspec": {
"display_name": "Python 3",
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python"
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.12"
},
"widgets": {
"application/vnd.jupyter.widget-state+json": {
Expand Down Expand Up @@ -2225,5 +2235,5 @@
}
},
"nbformat": 4,
"nbformat_minor": 0
"nbformat_minor": 1
}
18 changes: 14 additions & 4 deletions examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_E5.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,7 +66,7 @@
}
],
"source": [
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum"
"!pip install -q --upgrade transformers[onnx]==4.29.1 optimum tensorflow"
]
},
{
Expand Down Expand Up @@ -376,13 +376,23 @@
"provenance": []
},
"kernelspec": {
"display_name": "Python 3",
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python"
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.12"
}
},
"nbformat": 4,
"nbformat_minor": 0
"nbformat_minor": 1
}
18 changes: 14 additions & 4 deletions examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_MPNet.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,7 +66,7 @@
}
],
"source": [
"!pip install -q --upgrade transformers[onnx]==4.33.1 optimum"
"!pip install -q --upgrade transformers[onnx]==4.33.1 optimum tensorflow"
]
},
{
Expand Down Expand Up @@ -465,11 +465,21 @@
"toc_visible": true
},
"kernelspec": {
"display_name": "Python 3",
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python"
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.12"
},
"widgets": {
"application/vnd.jupyter.widget-state+json": {
Expand Down Expand Up @@ -2529,5 +2539,5 @@
}
},
"nbformat": 4,
"nbformat_minor": 0
"nbformat_minor": 1
}
4 changes: 2 additions & 2 deletions examples/python/transformers/onnx/HuggingFace_ONNX_in_Spark_NLP_Whisper.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@
}
],
"source": [
"!pip install -q --upgrade \"transformers[onnx]==4.31.0\" optimum"
"!pip install -q --upgrade \"transformers[onnx]==4.31.0\" optimum tensorflow"
]
},
{
Expand Down Expand Up @@ -550,4 +550,4 @@
},
"nbformat": 4,
"nbformat_minor": 0
}
}
5 changes: 3 additions & 2 deletions src/main/scala/com/johnsnowlabs/ml/ai/Albert.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.ai.util.PrepareEmbeddings
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sentencepiece.{SentencePieceWrapper, SentencepieceEncoder}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
Expand Down Expand Up @@ -83,6 +83,7 @@ private[johnsnowlabs] class Albert(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

// keys representing the input and output tensors of the ALBERT model
private val SentenceStartTokenId = spp.getSppModel.pieceToId("[CLS]")
Expand All @@ -107,7 +108,7 @@ private[johnsnowlabs] class Albert(

case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
7 changes: 4 additions & 3 deletions src/main/scala/com/johnsnowlabs/ml/ai/AlbertClassification.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sentencepiece.{SentencePieceWrapper, SentencepieceEncoder}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
Expand Down Expand Up @@ -56,6 +56,7 @@ private[johnsnowlabs] class AlbertClassification(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

// keys representing the input and output tensors of the ALBERT model
protected val sentencePadTokenId: Int = spp.getSppModel.pieceToId("[pad]")
Expand Down Expand Up @@ -210,7 +211,7 @@ private[johnsnowlabs] class AlbertClassification(
val output = if (sequence) "logits" else "last_hidden_state"

// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down Expand Up @@ -348,7 +349,7 @@ private[johnsnowlabs] class AlbertClassification(
batch: Seq[Array[Int]],
maxSentenceLength: Int): (Array[Float], Array[Float]) = {
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
7 changes: 4 additions & 3 deletions src/main/scala/com/johnsnowlabs/ml/ai/Bert.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.ai.util.PrepareEmbeddings
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
import com.johnsnowlabs.ml.util.{ModelArch, ONNX, TensorFlow}
Expand Down Expand Up @@ -66,6 +66,7 @@ private[johnsnowlabs] class Bert(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

private def sessionWarmup(): Unit = {
val dummyInput =
Expand All @@ -90,7 +91,7 @@ private[johnsnowlabs] class Bert(

case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down Expand Up @@ -193,7 +194,7 @@ private[johnsnowlabs] class Bert(
val embeddings = detectedEngine match {
case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
7 changes: 4 additions & 3 deletions src/main/scala/com/johnsnowlabs/ml/ai/BertClassification.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
import com.johnsnowlabs.ml.util.{ONNX, TensorFlow}
Expand Down Expand Up @@ -62,6 +62,7 @@ private[johnsnowlabs] class BertClassification(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

def tokenizeWithAlignment(
sentences: Seq[TokenizedSentence],
Expand Down Expand Up @@ -222,7 +223,7 @@ private[johnsnowlabs] class BertClassification(
maxSentenceLength: Int): Array[Float] = {

// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down Expand Up @@ -452,7 +453,7 @@ private[johnsnowlabs] class BertClassification(
batch: Seq[Array[Int]],
maxSentenceLength: Int): (Array[Float], Array[Float]) = {
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
5 changes: 3 additions & 2 deletions src/main/scala/com/johnsnowlabs/ml/ai/CamemBert.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.ai.util.PrepareEmbeddings
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sentencepiece.{SentencePieceWrapper, SentencepieceEncoder}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
Expand Down Expand Up @@ -54,6 +54,7 @@ private[johnsnowlabs] class CamemBert(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

/** HACK: These tokens were added by fairseq but don't seem to be actually used when duplicated
* in the actual # sentencepiece vocabulary (this is the case for '''<s>''' and '''</s>''')
Expand Down Expand Up @@ -82,7 +83,7 @@ private[johnsnowlabs] class CamemBert(

case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
5 changes: 3 additions & 2 deletions src/main/scala/com/johnsnowlabs/ml/ai/DeBerta.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.ai.util.PrepareEmbeddings
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sentencepiece._
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
Expand Down Expand Up @@ -52,6 +52,7 @@ class DeBerta(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

// keys representing the input and output tensors of the DeBERTa model
private val SentenceStartTokenId = spp.getSppModel.pieceToId("[CLS]")
Expand All @@ -68,7 +69,7 @@ class DeBerta(

case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
5 changes: 3 additions & 2 deletions src/main/scala/com/johnsnowlabs/ml/ai/DistilBert.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ package com.johnsnowlabs.ml.ai

import ai.onnxruntime.OnnxTensor
import com.johnsnowlabs.ml.ai.util.PrepareEmbeddings
import com.johnsnowlabs.ml.onnx.OnnxWrapper
import com.johnsnowlabs.ml.onnx.{OnnxSession, OnnxWrapper}
import com.johnsnowlabs.ml.tensorflow.sign.{ModelSignatureConstants, ModelSignatureManager}
import com.johnsnowlabs.ml.tensorflow.{TensorResources, TensorflowWrapper}
import com.johnsnowlabs.ml.util.{ModelArch, ONNX, TensorFlow}
Expand Down Expand Up @@ -82,6 +82,7 @@ private[johnsnowlabs] class DistilBert(
if (tensorflowWrapper.isDefined) TensorFlow.name
else if (onnxWrapper.isDefined) ONNX.name
else TensorFlow.name
private val onnxSessionOptions: Map[String, String] = new OnnxSession().getSessionOptions

private def sessionWarmup(): Unit = {
val dummyInput =
Expand All @@ -103,7 +104,7 @@ private[johnsnowlabs] class DistilBert(
val embeddings = detectedEngine match {
case ONNX.name =>
// [nb of encoded sentences , maxSentenceLength]
val (runner, env) = onnxWrapper.get.getSession()
val (runner, env) = onnxWrapper.get.getSession(onnxSessionOptions)

val tokenTensors =
OnnxTensor.createTensor(env, batch.map(x => x.map(x => x.toLong)).toArray)
Expand Down
Loading
Loading