torchrec
torchrec copied to clipboard
pytorch
Pytorch domain library for recommendation systems
Hi TorchRec Team, We've recently been trying to use the dynamic embedding feature in torchrec contrib, but we have encountered a few challenges. The process may result in a segmentation...
Summary: More of an RFC diff. depends on https://github.com/pytorch/torchrec/pull/1837 (see diagram there) The high level idea is we want to disagg the dense and sparse tower placement in rec model...
Differential Revision: D56915458 Privacy Context Container: 1203980333745195
Summary: The first fused parameter will override the other tbes. Fixing by using config.is_weighted Reviewed By: hlin09 Differential Revision: D56847642
Summary: design doc: https://docs.google.com/document/d/17_nqdEtH6B_ev9Gnuw2mpgtFq4dqzC6-XUdw18R4F8Q/ # changes ## in ITEP make remap a callback of input dist ## in torchrec copy callbacks from input dist to fused input dist # impact...
Summary: Adding dynamo coverage for KJT methods: - permute - split - regroup_as_dict - getitem - todict split and getitem tests need additional checks (similar to pre slice check). Extracted...
Summary: Fix pyre-check, ex: https://github.com/pytorch/torchrec/actions/runs/9036052208/job/24832104332 Reviewed By: henrylhtsang Differential Revision: D57224368
These talks capture what model parallelism is, what we are trying to do, and different components of TorchRec [Intro to Model Parallelism for Recommendations Systems](https://docs.google.com/presentation/d/1RJAoYGbqgyAPCGs8eKOBDhQTqD9MrznlaIOIC_f_3ww/edit?usp=drive_link) by Dennis [Torchrec Part 1](https://docs.google.com/presentation/d/1C6WKsbQBD8HIORH8bts-T5WK97dVidDNbHZZW738SSU/edit?usp=drive_link)...
Summary: Make KJT permute work in edge cases for torch.export: https://fb.workplace.com/groups/6829516587176185/posts/7206415192819654/?comment_id=7207715762689597&reply_comment_id=7210658785728628. A cleaner solution than before as well. Necessary for enabling PT2 eager model processing on AIMP Differential Revision: D57162331