viswa-nvidia

Results 163 comments of viswa-nvidia

[BUG]Flaky unit tests for models

@rnyak , I see this is blocked. Why is this blocked ?

[BUG] The negative sampling for ranking gives AUC 0.00 value when we add `sampling` layer to the Model class

@rnyak , please triage this bug

[Task] Update notebooks to note we require TF 2.8+ in the notebooks.

@benfred , do we need to track this still?

[TASK] Support shared embeddings across the towers

check with @rnyak and split the ticket

[BUG] get_aliccp() failed while processing Ali CCP dataset

Setting P1 as there is a workaround available.

[BUG] WDL training notebook for HugeCTR processing workflow fails with TypeError

@Spartee , Thank you for posting this bug. We have moved this ticket to the HugeCTR repository as this bug is related to the HugeCTR notebook.

[Task] Transformers4Rec Serving using Merlin Schema

Moving this to T4R repo.

[Task] Add Example for Transformer4Rec Serving Example

We ahve an example with python backend. This example with torch script.

[BUG] Dataloader doesnt release memory and memory growth

@benfred , please check if there is a fix from @oliverholworthy and link it. Thank you

[Task] Profiling the training&evaluation pipeline for retrieval and ranking models

@gabrielspmoreira , is this a bug or a feature request. The description says steps to reproduce bug. Please clarify.