Automatic Optimization of Facebook’s Deep Learning Recommendation Model
This repository includes the source code of Reinforcement Learning approach to automatically design Facebook's Deep Learning recommendation Model (DLRM) implemented by PyTorch.
Diffusion AutoRL-DLRM (master)
Recent Commits
Recent Commits
| Commit | Author | Details | Committed | ||||
|---|---|---|---|---|---|---|---|
| df91e679e1f8 | arman | Cleaned | Feb 5 | ||||
| 162d92509dec | arman | Cleaned and comments added | Feb 2 | ||||
| 5864fe74b2e0 | Francloz | RL | Aug 29 2020 | ||||
| 62b35f990129 | Francloz | Only RL polished | Aug 29 2020 | ||||
| 7c6bd51bfbd3 | Francloz | All versions | Aug 29 2020 | ||||
| 124542f7de1a | Francloz | New prefetcher, better functionality, optimized dataset loading | Mar 3 2020 | ||||
| 3d5bd409efe3 | Francloz | Parallelized across samples. | Feb 9 2020 | ||||
| cb7c92bc0843 | Francloz | Parallelized model using replicates. | Feb 9 2020 | ||||
| 221f878b606b | Francloz | New unified tensor version. Easy parallelization. | Feb 9 2020 | ||||
| 1c9f4f949a29 | Francloz | Refactoring. Simplified application of embedding bag | Feb 8 2020 | ||||
| c80087e264e9 | Francloz | Refactoring. Simplification of parallel forward. | Feb 8 2020 | ||||
| b66f53ef71a2 | Francloz | Bug fix | Feb 7 2020 | ||||
| d5191b0760bf | Francloz | First commit. Original and optimized versions side by side. | Feb 5 2020 | ||||
| bdba86d8df19 | mnaumovfb | Adding support for binary loader proposed in pull request https://github. | Jan 26 2020 | ||||
| 7c55962facb4 | mnaumovfb | Fixing a bug in dataset loader. When multiple passes over the data are done you… | Jan 22 2020 |
c4science · Help