TorchEasyRec implements state of the art deep learning models used in common recommendation tasks: candidate generation(matching), scoring(ranking), and multi-task learning. It improves the efficiency of generating high performance models by simple configuration and easy customization.
- Local / PAI-DLC / PAI-DSW / EMR-DataScience
- MaxCompute Table
- OSS files
- CSV files
- Parquet files
- Flexible feature config and model config
- Easy to implement customized models
- Easy deployment to EAS: automatic scaling, easy monitoring
- Efficient and robust feature generation
- Large scale embedding with different sharding strategies
- Hybrid data-parallelism/model-parallelism
- Optimized kernels for RecSys powered by TorchRec
- Consistency guarantee: train and serving
- IdFeature / RawFeature / ComboFeature / LookupFeature / MatchFeature / ExprFeature / OverlapFeature / TokenizeFeature / SequenceIdFeature / SequenceRawFeature / SequenceFeature
- DSSM / TDM
- DeepFM / MultiTower / DIN
- MMoE / DBMTL / PLE
- More models in development
Any contributions you make are greatly appreciated!
- Please report bugs by submitting a issue.
- Please submit contributions using pull requests.
- Please refer to the Development document for more details.
- DingDing Group: 32260796. (TorchEasyRec usage general discussion.)
- DingDing Group2: 37930014162, click this url or scan QrCode to join
- Email Group: [email protected].
- If you have any questions about how to use TorchEasyRec, please join the DingTalk group and contact us.
- If you have enterprise service needs or need to purchase Alibaba Cloud services to build a recommendation system, please join the DingTalk group to contact us.
TorchEasyRec is released under Apache License 2.0. Please note that third-party libraries may not have the same license as TorchEasyRec.
