recommenders

recommenders copied to clipboard

fix keras compatibility bug in factorized_top_k

13

The bug is summarized by the following issue comment. https://github.com/tensorflow/recommenders/issues/712#issuecomment-2041163592

getumen

Very low topk %, real-world accuracy around 40% at k=5000

1

Sorry if this may be the wrong place to ask, I couldn't find a lead anywhere after weeks of googling. I've implemented a retrieval model to create embeddings for cosine...

pthieu

load_weights can't work

1

session 1. def build_model(): model = RetrievalModel(item_model, user_model) learning_rate = 0.01 model.compile(optimizer=tf.keras.optimizers.Adagrad(learning_rate)) return model model = build_model() model.fit(behavior_dataset, epochs=30) model.save_weights(save_path,overwrite=True ) compute_loss_args = { "user_id" : tf.constant(["45"]), "work_id" : tf.constant(["45"]),...

mahaidongs

[Question] two-tower-model + infoNCE how to optimize

1

I have tried a two-tower model (user and query) in a real industrial scenario using contrastive learning. The samples are all actual click samples, and the loss function is InfoNCE....

unshaven

feat: Extend 'sample_listwise' to support more movielens Dataset features

3

The current function `sample_listwise` supports a limited number of features that are hardcoded in the function definition. This pull requests makes it easy for other features to be used as...

keivanipchihagh

Hi, as the title says, the performance of the model drops when I use a cluster of GPUs. The (custom) training is being done in vertex training service. This is...

hugoferrero

How can I use sample_weights for retrieval task?

1

Hi, I'm trying to build a retrieval model. I have positive samples and some weights corresponding to each of these sample which helps to denote that for the given sample...

modic23

Please Help ValueError: Cannot convert '('c', 'o', 'u', 'n', 't', 'e', 'r')' to a shape. Found invalid entry 'c' of type '<class 'str'>'.

3

Hi, I have been working through the Recommending movies: retrieval tutorial and in the Metrics portion of the jupyter notebook I have been receiving a ValueError: Cannot convert '('c', 'o',...

Levine-Sam

Getting below error when trying Quickstart example. --------------- > File "/Users/terry/Documents/project/recommender/recommenders/collab_filtering_movie.py", line 245, in > index.index_from_dataset( > File "/Users/terry/Documents/project/recommender/recommenders/tensorflow_recommenders/layers/factorized_top_k.py", line 215, in index_from_dataset > return self.index(candidates, identifiers) > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ >...

flynhigher

Fix for ValueError: The shape of the target variable and the shape of the target value in variable.assign(value) must match

1

Created an issue for this error - #726

flynhigher