Databricks Runtime 10.3 für ML (nicht unterstützt)

Databricks Runtime 10.3 für Machine Learning bietet eine sofort einsatzbereite Umgebung für maschinelles Lernen und Data Science auf Basis von Databricks Runtime 10.3 (nicht unterstützt). Databricks Runtime ML enthält viele beliebte Machine Learning-Bibliotheken, einschließlich TensorFlow, PyTorch und XGBoost. Databricks Runtime ML enthält AutoML, ein Tool zum automatischen Trainieren von Machine Learning-Pipelines. Databricks Runtime ML unterstützt auch verteiltes Deep Learning-Training mit Horovod.

Weitere Informationen, einschließlich Anweisungen zum Erstellen eines Databricks Runtime ML-Clusters, finden Sie unter KI und Machine Learning in Databricks.

Neue Features und Verbesserungen

Databricks Runtime 10.3 ML basiert auf Databricks Runtime 10.3. Informationen zu den Neuerungen in Databricks Runtime 10.3 (einschließlich Apache Spark MLlib und SparkR) finden Sie in den Versionshinweisen zu Databricks Runtime 10.3 (nicht unterstützt).

Verbesserungen an Databricks AutoML

Die folgenden Verbesserungen wurden an Databricks AutoML vorgenommen.

AutoML unterstützt jetzt das ARIMA-Modell für Prognosen

Zusätzlich zu Prophet erstellt und wertet AutoML jetzt auch ARIMA-Modelle für Prognoseprobleme aus.

Ausschließen von Spalten aus einem Dataset

Wenn Sie die AutoML-API verwenden, können Sie Spalten angeben, die AutoML bei seinen Berechnungen ignorieren soll. Dies ist nur für Klassifizierungs- und Regressionsprobleme verfügbar. Weitere Informationen finden Sie unter Klassifizierungs- und Regressionsparameter.

Ausschließen von Algorithmusframeworks von einer AutoML-Ausführung

Sie können Algorithmus-Frameworks wie scikit-learn angeben, die AutoML bei der Entwicklung von Modellen nicht berücksichtigen soll. Weitere Informationen finden Sie unter Erweiterte Konfigurationen und Klassifizierungs- und Regressionsparameter.

max_trials veraltet

Der Parameter max_trials ist veraltet und wird aus dem nächsten Hauptrelease von Databricks Runtime ML entfernt. Verwenden Sie timeout_minutes, um die Dauer einer AutoML-Ausführung zu steuern. Ab Databricks Runtime 10.1 ML verfügt AutoML außerdem über eine Funktion zum frühzeitigen Beenden des Trainings und der Optimierung von Modellen, wenn sich die Validierungsmetrik nicht mehr verbessert.

Verbesserungen am Featurespeicher von Databricks

Sie können jetzt zeitpunktbezogene Lookups auf Zeitreihen-Featuretabellen anwenden. Informationen finden Sie unter Verwenden von Zeitreihen-Featuretabellen mit Point-in-Time-Support.

Automatische Databricks-Protokollierung (allgemeine Verfügbarkeit)

Die automatische Databricks-Protokollierung ist jetzt in Databricks Runtime 10.3 ML allgemein verfügbar. Die automatische Databricks-Protokollierung ist eine Lösung ohne Programmieraufwand, die eine automatische Nachverfolgung von Experimenten für Machine Learning-Trainingssitzungen in Azure Databricks ermöglicht. Mithilfe der automatischen Databricks-Protokollierung werden Modellparameter, Metriken, Dateien und Informationen zur Herkunft automatisch erfasst, wenn Sie Modelle anhand einer Vielzahl beliebter Machine Learning-Bibliotheken trainieren. Trainingssitzungen werden als MLflow-Nachverfolgungsausführungen aufgezeichnet. Modelldateien werden auch nachverfolgt, sodass Sie sie problemlos in der MLflow-Modellregistrierung protokollieren und für die Bewertung in Echtzeit mithilfe der MLflow-Modellbereitstellung bereitstellen können.

Weitere Informationen finden Sie unter Automatische Databricks-Protokollierung.

Systemumgebung

Die Systemumgebung der Databricks Runtime 10.3 ML unterscheidet sich wie folgt von der Databricks Runtime 10.3:

Bibliotheken

In den folgenden Abschnitten sind die Bibliotheken aufgelistet, die in Databricks Runtime 10.3 ML enthalten sind und sich von den in Databricks Runtime 10.3 enthaltenen Bibliotheken unterscheiden.

Inhalt dieses Abschnitts:

Bibliotheken der obersten Ebene

Databricks Runtime 10.3 ML enthält die folgenden Bibliotheken der obersten Ebene:

Python-Bibliotheken

Databricks Runtime 10.3 ML verwendet Virtualenv zur Verwaltung von Python-Paketen und enthält viele beliebte ML-Pakete.

Zusätzlich zu den Paketen, die in den folgenden Abschnitten aufgeführt sind, enthält Databricks Runtime 10.3 ML auch die folgenden Pakete:

  • hyperopt 0.2.7.db1
  • sparkdl 2.2.0-db5
  • feature_store 0.3.7
  • automl 1.6.0

Python-Bibliotheken in CPU-Clustern

Bibliothek Version Bibliothek Version Bibliothek Version
absl-py 0.11.0 Antergos Linux 2015.10 (ISO-Rolling) appdirs 1.4.4
argon2-cffi 20.1.0 astor 0.8.1 astunparse 1.6.3
async-generator 1.10 attrs 20.3.0 backcall 0.2.0
bcrypt 3.2.0 bidict 0.21.4 bleach 3.3.0
blis 0.7.4 boto3 1.16.7 botocore 1.19.7
cachetools 4.2.4 Katalog 2.0.6 certifi 2020.12.5
cffi 1.14.5 chardet 4.0.0 Klicken 7.1.2
cloudpickle 1.6.0 cmdstanpy 0.9.68 configparser 5.0.1
convertdate 2.3.2 cryptography 3.4.7 cycler 0.10.0
cymem 2.0.5 Cython 0.29.23 databricks-automl-runtime 0.2.5
databricks-cli 0.16.2 dbl-tempo 0.1.2 dbus-python 1.2.16
decorator 5.0.6 defusedxml 0.7.1 dill 0.3.2
diskcache 5.2.1 distlib 0.3.4 distro-info 0.23ubuntu1
entrypoints 0,3 ephem 4.1.3 facets-overview 1.0.0
fasttext 0.9.2 filelock 3.0.12 Flask 1.1.2
flatbuffers 2.0 fsspec 0.9.0 future 0.18.2
gast 0.4.0 gitdb 4.0.7 GitPython 3.1.12
google-auth 1.22.1 google-auth-oauthlib 0.4.2 google-pasta 0.2.0
grpcio 1.39.0 gunicorn 20.0.4 gviz-api 1.10.0
h5py 3.1.0 hijri-converter 2.2.2 holidays 0,12
horovod 0.23.0 htmlmin 0.1.12 huggingface-hub 0.1.2
idna 2.10 ImageHash 4.2.1 imbalanced-learn 0.8.1
importlib-metadata 3.10.0 ipykernel 5.3.4 ipython 7.22.0
ipython-genutils 0.2.0 ipywidgets 7.6.3 isodate 0.6.0
itsdangerous 1.1.0 jedi 0.17.2 Jinja2 2.11.3
jmespath 0.10.0 joblib 1.0.1 joblibspark 0.3.0
jsonschema 3.2.0 jupyter-client 6.1.12 jupyter-core 4.7.1
jupyterlab-pygments 0.1.2 jupyterlab-widgets 1.0.0 keras 2.7.0
Keras-Preprocessing 1.1.2 kiwisolver 1.3.1 koalas 1.8.2
korean-lunar-calendar 0.2.1 langcodes 3.3.0 libclang 12.0.0
lightgbm 3.3.1 llvmlite 0.38.0 LunarCalendar 0.0.9
Mako 1.1.3 Markdown 3.3.3 MarkupSafe 2.0.1
matplotlib 3.4.2 missingno 0.5.0 mistune 0.8.4
mleap 0.18.1 mlflow-skinny 1.23.0 multimethod 1.6
murmurhash 1.0.5 nbclient 0.5.3 nbconvert 6.0.7
nbformat 5.1.3 nest-asyncio 1.5.1 networkx 2.5
nltk 3.6.1 Notebook 6.3.0 numba 0.55.0
numpy 1.20.1 oauthlib 3.1.0 opt-einsum 3.3.0
Packen 21,3 Pandas 1.2.4 pandas-profiling 3.1.0
pandocfilters 1.4.3 paramiko 2.7.2 parso 0.7.0
pathy 0.6.0 patsy 0.5.1 petastorm 0.11.3
pexpect 4.8.0 phik 0.12.0 pickleshare 0.7.5
Pillow 8.2.0 pip 21.0.1 plotly 5.5.0
pmdarima 1.8.4 preshed 3.0.5 prometheus-client 0.10.1
prompt-toolkit 3.0.17 prophet 1.0.1 protobuf 3.17.2
psutil 5.8.0 psycopg2 2.8.5 ptyprocess 0.7.0
pyarrow 4.0.0 pyasn1 0.4.8 pyasn1-modules 0.2.8
pybind11 2.9.0 pycparser 2,20 pydantic 1.8.2
Pygments 2.8.1 PyGObject 3.36.0 PyMeeus 0.5.11
PyNaCl 1.4.0 pyodbc 4.0.30 pyparsing 2.4.7
pyrsistent 0.17.3 pystan 2.19.1.1 python-apt 2.0.0+ubuntu0.20.4.6
Python-dateutil 2.8.1 python-editor 1.0.4 python-engineio 4.3.0
python-socketio 5.4.1 pytz 2020.5 PyWavelets 1.1.1
PyYAML 5.4.1 pyzmq 20.0.0 regex 2021.4.4
requests 2.25.1 requests-oauthlib 1.3.0 requests-unixsocket 0.2.0
rsa 4.7.2 s3transfer 0.3.7 sacremoses 0.0.46
scikit-learn 0.24.1 scipy 1.6.2 seaborn 0.11.1
Send2Trash 1.5.0 setuptools 52.0.0 setuptools-git 1.2
shap 0.40.0 simplejson 3.17.2 sechs 1.15.0
slicer 0.0.7 smart-open 5.2.0 smmap 3.0.5
spacy 3.2.1 spacy-legacy 3.0.8 spacy-loggers 1.0.1
spark-tensorflow-distributor 1.0.0 sqlparse 0.4.1 srsly 2.4.1
ssh-import-id 5.10 statsmodels 0.12.2 tabulate 0.8.7
tangled-up-in-unicode 0.1.0 tenacity 6.2.0 tensorboard 2.7.0
tensorboard-data-server 0.6.1 tensorboard-plugin-profile 2.5.0 tensorboard-plugin-wit 1.8.1
tensorflow-cpu 2.7.0 tensorflow-estimator 2.7.0 tensorflow-io-gcs-filesystem 0.23.1
termcolor 1.1.0 terminado 0.9.4 testpath 0.4.4
thinc 8.0.12 threadpoolctl 2.1.0 tokenizers 0.10.3
torch 1.10.1+cpu torchvision 0.11.2+cpu tornado 6.1
tqdm 4.59.0 traitlets 5.0.5 transformers 4.15.0
typer 0.3.2 typing-extensions 3.7.4.3 ujson 4.0.2
unattended-upgrades 0,1 urllib3 1.25.11 virtualenv 20.4.1
Visionen 0.7.4 Wasabi 0.8.2 wcwidth 0.2.5
webencodings 0.5.1 websocket-client 0.57.0 Werkzeug 1.0.1
wheel 0.36.2 widgetsnbextension 3.5.1 wrapt 1.12.1
xgboost 1.5.1 zipp 3.4.1

Python-Bibliotheken für GPU-Cluster

Bibliothek Version Bibliothek Version Bibliothek Version
absl-py 0.11.0 Antergos Linux 2015.10 (ISO-Rolling) appdirs 1.4.4
argon2-cffi 20.1.0 astor 0.8.1 astunparse 1.6.3
async-generator 1.10 attrs 20.3.0 backcall 0.2.0
bcrypt 3.2.0 bidict 0.21.4 bleach 3.3.0
blis 0.7.4 boto3 1.16.7 botocore 1.19.7
cachetools 4.2.4 Katalog 2.0.6 certifi 2020.12.5
cffi 1.14.5 chardet 4.0.0 Klicken 7.1.2
cloudpickle 1.6.0 cmdstanpy 0.9.68 configparser 5.0.1
convertdate 2.3.2 cryptography 3.4.7 cycler 0.10.0
cymem 2.0.5 Cython 0.29.23 databricks-automl-runtime 0.2.5
databricks-cli 0.16.2 dbl-tempo 0.1.2 dbus-python 1.2.16
decorator 5.0.6 defusedxml 0.7.1 dill 0.3.2
diskcache 5.2.1 distlib 0.3.4 distro-info 0.23ubuntu1
entrypoints 0,3 ephem 4.1.3 facets-overview 1.0.0
fasttext 0.9.2 filelock 3.0.12 Flask 1.1.2
flatbuffers 2.0 fsspec 0.9.0 future 0.18.2
gast 0.4.0 gitdb 4.0.7 GitPython 3.1.12
google-auth 1.22.1 google-auth-oauthlib 0.4.2 google-pasta 0.2.0
grpcio 1.39.0 gunicorn 20.0.4 gviz-api 1.10.0
h5py 3.1.0 hijri-converter 2.2.2 holidays 0,12
horovod 0.23.0 htmlmin 0.1.12 huggingface-hub 0.1.2
idna 2.10 ImageHash 4.2.1 imbalanced-learn 0.8.1
importlib-metadata 3.10.0 ipykernel 5.3.4 ipython 7.22.0
ipython-genutils 0.2.0 ipywidgets 7.6.3 isodate 0.6.0
itsdangerous 1.1.0 jedi 0.17.2 Jinja2 2.11.3
jmespath 0.10.0 joblib 1.0.1 joblibspark 0.3.0
jsonschema 3.2.0 jupyter-client 6.1.12 jupyter-core 4.7.1
jupyterlab-pygments 0.1.2 jupyterlab-widgets 1.0.0 keras 2.7.0
Keras-Preprocessing 1.1.2 kiwisolver 1.3.1 koalas 1.8.2
korean-lunar-calendar 0.2.1 langcodes 3.3.0 libclang 12.0.0
lightgbm 3.3.1 llvmlite 0.38.0 LunarCalendar 0.0.9
Mako 1.1.3 Markdown 3.3.3 MarkupSafe 2.0.1
matplotlib 3.4.2 missingno 0.5.0 mistune 0.8.4
mleap 0.18.1 mlflow-skinny 1.23.0 multimethod 1.6
murmurhash 1.0.5 nbclient 0.5.3 nbconvert 6.0.7
nbformat 5.1.3 nest-asyncio 1.5.1 networkx 2.5
nltk 3.6.1 Notebook 6.3.0 numba 0.55.0
numpy 1.20.1 oauthlib 3.1.0 opt-einsum 3.3.0
Packen 21,3 Pandas 1.2.4 pandas-profiling 3.1.0
pandocfilters 1.4.3 paramiko 2.7.2 parso 0.7.0
pathy 0.6.0 patsy 0.5.1 petastorm 0.11.3
pexpect 4.8.0 phik 0.12.0 pickleshare 0.7.5
Pillow 8.2.0 pip 21.0.1 plotly 5.5.0
pmdarima 1.8.4 preshed 3.0.5 prompt-toolkit 3.0.17
prophet 1.0.1 protobuf 3.17.2 psutil 5.8.0
psycopg2 2.8.5 ptyprocess 0.7.0 pyarrow 4.0.0
pyasn1 0.4.8 pyasn1-modules 0.2.8 pybind11 2.9.0
pycparser 2,20 pydantic 1.8.2 Pygments 2.8.1
PyGObject 3.36.0 PyMeeus 0.5.11 PyNaCl 1.4.0
pyodbc 4.0.30 pyparsing 2.4.7 pyrsistent 0.17.3
pystan 2.19.1.1 python-apt 2.0.0+ubuntu0.20.4.6 Python-dateutil 2.8.1
python-editor 1.0.4 python-engineio 4.3.0 python-socketio 5.4.1
pytz 2020.5 PyWavelets 1.1.1 PyYAML 5.4.1
pyzmq 20.0.0 regex 2021.4.4 requests 2.25.1
requests-oauthlib 1.3.0 requests-unixsocket 0.2.0 rsa 4.7.2
s3transfer 0.3.7 sacremoses 0.0.46 scikit-learn 0.24.1
scipy 1.6.2 seaborn 0.11.1 Send2Trash 1.5.0
setuptools 52.0.0 setuptools-git 1.2 shap 0.40.0
simplejson 3.17.2 sechs 1.15.0 slicer 0.0.7
smart-open 5.2.0 smmap 3.0.5 spacy 3.2.1
spacy-legacy 3.0.8 spacy-loggers 1.0.1 spark-tensorflow-distributor 1.0.0
sqlparse 0.4.1 srsly 2.4.1 ssh-import-id 5.10
statsmodels 0.12.2 tabulate 0.8.7 tangled-up-in-unicode 0.1.0
tenacity 6.2.0 tensorboard 2.7.0 tensorboard-data-server 0.6.1
tensorboard-plugin-profile 2.5.0 tensorboard-plugin-wit 1.8.1 tensorflow 2.7.0
tensorflow-estimator 2.7.0 tensorflow-io-gcs-filesystem 0.23.1 termcolor 1.1.0
terminado 0.9.4 testpath 0.4.4 thinc 8.0.12
threadpoolctl 2.1.0 tokenizers 0.10.3 torch 1.10.1+cu111
torchvision 0.11.2+cu111 tornado 6.1 tqdm 4.59.0
traitlets 5.0.5 transformers 4.15.0 typer 0.3.2
typing-extensions 3.7.4.3 ujson 4.0.2 unattended-upgrades 0,1
urllib3 1.25.11 virtualenv 20.4.1 Visionen 0.7.4
Wasabi 0.8.2 wcwidth 0.2.5 webencodings 0.5.1
websocket-client 0.57.0 Werkzeug 1.0.1 wheel 0.36.2
widgetsnbextension 3.5.1 wrapt 1.12.1 xgboost 1.5.1
zipp 3.4.1

Spark-Pakete mit Python-Modulen

Spark-Paket Python-Modul Version
graphframes graphframes 0.8.2-db1-spark3.2

R-Bibliotheken

Die R-Bibliotheken sind mit den R-Bibliotheken in Databricks Runtime 10.3 identisch.

Java- und Scala-Bibliotheken (Scala 2.12-Cluster)

Zusätzlich zu Java- und Scala-Bibliotheken in Databricks Runtime 10.3 enthält Databricks Runtime 10.3 ML die folgenden JAR-Dateien:

CPU-Cluster

Gruppen-ID Artefakt-ID Version
com.typesafe.akka akka-actor_2.12 2.5.23
ml.combust.mleap mleap-databricks-runtime_2.12 0.18.1-23eb1ef
ml.dmlc xgboost4j-spark_2.12 1.5.1
ml.dmlc xgboost4j_2.12 1.5.1
org.graphframes graphframes_2.12 0.8.2-db1-spark3.2
org.mlflow mlflow-client 1.23.0
org.mlflow mlflow-spark 1.23.0
org.scala-lang.modules scala-java8-compat_2.12 0.8.0
org.tensorflow spark-tensorflow-connector_2.12 1.15.0

GPU-Cluster

Gruppen-ID Artefakt-ID Version
com.typesafe.akka akka-actor_2.12 2.5.23
ml.combust.mleap mleap-databricks-runtime_2.12 0.18.1-23eb1ef
ml.dmlc xgboost4j-spark_2.12 1.5.1
ml.dmlc xgboost4j_2.12 1.5.1
org.graphframes graphframes_2.12 0.8.2-db1-spark3.2
org.mlflow mlflow-client 1.23.0
org.mlflow mlflow-spark 1.23.0
org.scala-lang.modules scala-java8-compat_2.12 0.8.0
org.tensorflow spark-tensorflow-connector_2.12 1.15.0