mirror of https://github.com/hanzo-ml/trainer.git synced 2026-09-24 00:31:34 +00:00

Files

T

Ce GaoandGitHub 8082044198 feat(pytorch): Support elastic training (#1453 )

* chore: Update codegen and vendor

Signed-off-by: Ce Gao <[email protected]>

* fix: Update git config

Signed-off-by: Ce Gao <[email protected]>

* fix: Codegen

Signed-off-by: Ce Gao <[email protected]>

* feat: Add elastic examples

Signed-off-by: Ce Gao <[email protected]>

* chore: Update go mod

Signed-off-by: Ce Gao <[email protected]>

* chore: Update manifest

Signed-off-by: Ce Gao <[email protected]>

* feat: Implement elstic policy

Signed-off-by: Ce Gao <[email protected]>

* chore: Update vendor

Signed-off-by: Ce Gao <[email protected]>

* chore: Update swagger

Signed-off-by: Ce Gao <[email protected]>

* fix: Address comments

Signed-off-by: Ce Gao <[email protected]>

2021-11-26 06:35:12 +00:00

795 B

Raw Blame History

V1PyTorchJobSpec

PyTorchJobSpec is a desired state description of the PyTorchJob.

Properties

Name	Type	Description	Notes
elastic_policy	V1ElasticPolicy		[optional]
pytorch_replica_specs	dict(str, V1ReplicaSpec)	A map of PyTorchReplicaType (type) to ReplicaSpec (value). Specifies the PyTorch cluster configuration. For example, { "Master": PyTorchReplicaSpec, "Worker": PyTorchReplicaSpec, }
run_policy	V1RunPolicy

[Back to Model list] [Back to API list] [Back to README]

795 B Raw Blame History

V1PyTorchJobSpec

Properties

795 B

Raw Blame History