Pusher
E457349
Pusher is a TensorFlow Extended (TFX) component responsible for validating and deploying trained machine learning models to serving infrastructure.
Statements (46)
| Predicate | Object |
|---|---|
| instanceOf |
TFX component
ⓘ
software component ⓘ |
| configuredWith |
Pusher component arguments
ⓘ
PusherSpec ⓘ |
| consumes |
model blessing artifacts
ⓘ
trained model artifacts ⓘ |
| developedBy | Google NERFINISHED ⓘ |
| documentedAt | https://www.tensorflow.org/tfx/guide/pusher ⓘ |
| domain |
MLOps
NERFINISHED
ⓘ
machine learning engineering ⓘ |
| ensures |
only validated models are deployed
ⓘ
traceability of deployed models via metadata ⓘ |
| goal | make validated models available for serving ⓘ |
| hasFeature |
conditional model pushing based on validation
ⓘ
integration with ML Metadata ⓘ support for custom push destinations ⓘ versioned model deployment ⓘ |
| implementedIn | Python NERFINISHED ⓘ |
| inputFrom |
Evaluator
ⓘ
Trainer ⓘ |
| outputs |
deployment metadata
ⓘ
pushed model artifacts ⓘ |
| partOf |
TFX pipeline
ⓘ
TensorFlow Extended NERFINISHED ⓘ |
| platform | TensorFlow ecosystem NERFINISHED ⓘ |
| requires |
a blessed model from Evaluator
ⓘ
a trained model from Trainer ⓘ |
| responsibleFor |
checking model blessing status
ⓘ
exporting models for serving ⓘ model deployment ⓘ model validation ⓘ pushing models to serving infrastructure ⓘ |
| supports |
CI/CD style model rollout
ⓘ
TensorFlow Serving deployment ⓘ cloud storage deployment ⓘ custom deployment targets ⓘ local filesystem deployment ⓘ |
| usedIn |
ML production workflows
ⓘ
machine learning pipelines ⓘ |
| usedWith |
Evaluator
ⓘ
ExampleGen NERFINISHED ⓘ ExampleValidator NERFINISHED ⓘ SchemaGen NERFINISHED ⓘ StatisticsGen NERFINISHED ⓘ Trainer ⓘ Transform ⓘ |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.