Apdullah Yayık

Sorry, I just have noticed that `tf.distribute.MirroredStrategy()` succesfully works for LibRecommender models that is why I am closing this issue.

I realized that using `tf.distribute.MirroredStrategy()` within this library has a problem. That is most of the ops are done in GPU:0. How can I distribute works between GPUs fairly? 

By considering multi-gpu implementation on https://github.com/vahidk/EffectiveTensorflow/tree/v1#multi_gpu, for Wide and Deep Model with ranking task, I have added `make_parallel` to `WideDeep` class in `algorithms/wide_deep.py` ``` @staticmethod def make_parallel(fn, num_gpus, **kwargs): in_splits...

By the way, I wonder that in this library there is a limitation of users that have few interactions or not. Like discarding users that have only 1 or 2...

Another question: 1- In libreco library ,when evaluating a model of ranking task , do you make recommendation considering all available item profiles, at train, test, and eval sets, or...

This is my solution 1- I have added, `metric_name_monitored` to `print_metrics_ranking` function in `evaluation/evaluate.py` ```Python def print_metrics_ranking(metrics, y_prob=None, y_true=None, y_reco_list=None, y_true_list=None, users=None, k=10, train=True, metric_name_monitored=None): ``` 2- In `EvalMixin` class...

Thanks. For now, I am closing this issue. Note that whenever you notify me, I can send a PR.

So for this task, there are 3 todos like below: 1- `print_metrics_ranking` function has more than one role that are computing and printing metrics contrary to its name. I think...

The same error raises when using Cesar model even without multi_sparse features. As far as I see, user_vector and item_vector dont have OOV inside. Is there any suggestion? 

calling `assign_oov_vector` method exist at line 281 after the train loop, there is no missing line. The error is raised before calling `assign_oov_vector` at line 324 of `cesar.py`