DeepSpeed
a44fb58f - Fix AutoTP custom patterns: respect use_default_specs (#7827)

Commit
54 days ago
Fix AutoTP custom patterns: respect use_default_specs (#7827) The current code has the following issues: - `use_default_specs: false` doesn't work - Injection by the traditional pattern runs even when custom patterns are set - `mpu` needs to be passed to `deepspeed.initialize` (HF integration doesn't pass mpu) This PR fixes AutoTP setup to respect `use_default_specs: false` and disable the traditional injection path when custom patterns are enabled. Also, when `mpu` is not passed, we create a TP group in the initialization process. With these changes, the [related tests](https://github.com/deepspeedai/DeepSpeed/tree/master/tests/unit/model_parallelism) pass and [all AutoTP examples](https://github.com/tohtana/DeepSpeedExamples/tree/tohtana/custom_auto_tp/training/tensor_parallel) in DeepSpeedExamples work now ([PR](https://github.com/deepspeedai/DeepSpeedExamples/pull/998)). --------- Signed-off-by: Masahiro Tanaka <mtanaka@anyscale.com>
Author
Parents
Loading