alpa
Training and serving large-scale neural networks with auto parallelization.
APACHE-2.0 License
Bot releases are visible (Hide)
Maintenance release.
- Support manual sharding (#816)
- Optimize cross-mesh communication (#773, #798)
- Add publications (#885)
- Support new models: CodeGen, BLOOM-Z (#774, #844)
- Add priority scheduling in the model server (#852)
- Add guidance on strategy inspection (#879)
- Inference stage construction (#793, #799)
- Misc bug fixes (#876, #878, #873)
Published by merrymercy almost 2 years ago
- Rebase to jax v0.3.22 (#782)
- Support TPU backend for ShardParallel (#764)
- Misc improvements for LLM serving
Published by merrymercy about 2 years ago
- Fine-tune OPT with 3d parallelism #716
- Support bigscience/bloom #707
- Add model serving controller #720
- Jaxlib update
Published by merrymercy about 2 years ago
- Improve OPT generation speed (#661)
- Add logprobs endpoint (#679)
- Support global norm in pipeline parallelism (#672)
- Rebase jax to v0.3.15 (#682)
- Improve the stability of the server
Published by merrymercy about 2 years ago
- Improve OPT-175B demo https://opt.alpa.ai/. Our website has served 20k+ requests during the last week.
- Add OPT fine-tuning examples https://github.com/alpa-projects/alpa/tree/main/examples/opt_finetune
- Various bug fixes
Published by merrymercy about 2 years ago
- Improve OPT-175B demo https://opt.alpa.ai/
- Improve version management
- Various bug fixes
Published by merrymercy over 2 years ago
- [Feature] Serialize Parallel Plan (#587)
- [Test] Move test install to alpa package (#584)
- [CI/CD] New PyPI Infra (#577)
Published by merrymercy over 2 years ago
- [Tutorial] Serve OPT-175B (https://alpa-projects.github.io/tutorials/opt_serving.html)
- New website
- Better weight conversion scripts
- [API Change] Move layer construction decorators to ParallelMethod (#565)
Published by merrymercy over 2 years ago
- Interface change (https://github.com/alpa-projects/alpa/pull/540)
- Minor bug fixes
Published by zhisbug over 2 years ago
Major feature updates:
- Support serving Meta's open source OPT model 175B
- Distributed weight init
- Add a set of strategy-specific interfaces such as DataParallel, Zero2Parallel, Zero3Parallel
Minors:
- enhancement on the PyTorch frontend
- many bugfixes
- Doc update: Performance tuning guide, and a slide deck for Alpa
Published by zhisbug over 2 years ago
- PyTorch experimental support by @yf225
- small changes of interfaces on auto sharding
Published by merrymercy over 2 years ago
Published by zhisbug over 2 years ago
First release of Alpa v0.0.1
