search

tensorflow / recommenders-addons

Additional utils and helpers to extend TensorFlow when build recommendation systems, contributed and maintained by SIG Recommenders.
Apache License 2.0
592 stars 134 forks source link

Integrating tfrs with tfx #162

Closed yrianderreumaux closed 2 years ago

yrianderreumaux commented 3 years ago

I am trying to integrate a tfrs model with tfx, and I see that it's in the scope of recommender-addons ("End-to-end pipeline: how to train continuously, e.g. integrate with platforms like TFX"). However, I have not been able to find any docs or example pipelines to accomplish this. I am hoping to implement best practices, and in particular looking for help integrating the model with tft.

Any help would be much appreciated!

yrianderreumaux commented 3 years ago

Any updates on this?

Lifann commented 3 years ago

@Mr-Nineteen Would u plz help with the question?

yrianderreumaux commented 2 years ago

Checking in again on this request in case there are any updates.

Mr-Nineteen commented 2 years ago

we practice plan: TFRA on TensorFlow + Estimator + TFServing(TRT)，other components have not been tried.

  1. code demo: https://github.com/tensorflow/recommenders-addons/tree/master/demo/dynamic_embedding/movielens-100k-estimator
  2. Compatibility with Tensorflow Serving: https://github.com/tensorflow/recommenders-addons/blob/master/README.md#compatibility-with-tensorflow-serving
  3. If TFServing uses GPU, TRT uses tf.experimental.tensorrt.Converter(). At the same time, mlp is calculated on the GPU, and the embedding lookup is on the CPU, so that the performance is optimal.
yrianderreumaux commented 2 years ago

Thanks for this info @Mr-Nineteen!

yrianderreumaux commented 2 years ago

Closing until there are further updates specific to TFX