models

models copied to clipboard

[Task] Example notebooks for generation top_k recommendations with ranking and retrieval models

2

### Description Currently none of our example notebooks do not have any code script showcasing how one can do model prediction locally, generate top-K recommendations from the prediction results. This...

rnyak

Part of #669 ### Description Implement the save method on tensorflow models to write out both the tensorflow/keras artifacts as well as the merlin schema in a separate file.

oliverholworthy

[Task] Update evaluation in retrieval model examples

1

### Description Currently, 05-Retrieval-Model.ipynb doesnt evaluate the model (model.evaluate). I think we should extend the example by model.evaluate. 1. We should add a description, when using a validation dataset in...

bschifferer

### Description Updating all example notebooks to have a customized NVIDIA logo

bschifferer

Example of serving LightFM in Merlin Systems

2

viswa-nvidia

[BUG] Error when re-compile and fit a model

2

### Bug description Created a TwoTower Model or a Wide and Deep Model, firstly compile the model with the "adam" optiomizer and fit, and then call model.compile again with "sgd"...

mengyao00

[BUG] model.evaluate() gives much higher acc value for small batch size when we set `validation_data` int the model.fit()

4

### Bug description When we train Two-tower model and run model.evaluate() afterwards, we observe very different ndcg value wrt different batch size set in the `model.evaluate()`. if you remove `validation_data`...

rnyak

Top-k recommender model

3

Fixes #622 ### Goals :soccer: - [x] Decouple the top-k local prediction top-k evaluation from the retrieval contrastive learning task. - [ ] REfactor retrieval models and add a method...

sararb

### Bug description When training with (as best as I can tell) everything else held constant and the Dataset on the CPU, GPU memory consumption doesn't fluctuate during training. Throughout...

radekosmulski

[Task] Simplify Top-k recommender API

1

## Problem - Generating the top-k prediction from a trained the retrieval model currently requires the user to follow these steps (Not straightforward). Besides, the top-k recommender is a ModelBlock...

sararb