The TensorFlow runtime has components that are lazily initialized, which can cause high latency for the first request/s sent to a model after it is loaded. This latency can be several orders of magnitude higher than that of a single inference request.
To reduce the impact of lazy initialization on request latency, it's possible to trigger the initialization of the sub-systems and components at model load time by providing a sample set of inference requests along with the SavedModel. This process is known as "warming up" the model.
SavedModel Warmup is supported for Regress, Classify, MultiInference and Predict. To trigger warmup of the model at load time, attach a warmup data file under the assets.extra subfolder of the SavedModel directory.
Requirements for model warmup to work correctly:
- Warmup file name: 'tf_serving_warmup_requests'
- File location: assets.extra/
- File format: TFRecord with each record as a PredictionLog.
- Number of warmup records <= 1000.
- The warmup data must be representative of the inference requests used at serving.
Warm-up data generation
Warmup data can be added in two ways:
- By directly populating the warmup requests into your exported Saved Model.
This could be done by creating a script reading a list of sample
inference requests, converting each request into
(if it's originally in a different format) and using
to write the PredictionLog entries into
- By using TFX Infra Validator
option to export a Saved Model with warmup.
With this option the TFX Infa Validator will populate
YourSavedModel/assets.extra/tf_serving_warmup_requestsbased on the validation requests provided via RequestSpec.