tensor-fusion-helm-1.2.20

A Helm chart for Kubernetes

v1.27.0

1.27.0 (2025-05-16)

Features

• maintain all GPUs in memory instead of accessing the apiserver … (#186) (c3c1f5e)

v1.26.21

1.26.21 (2025-05-15)

Bug Fixes

• add annotation for tf workload, refine CRD schema, add gpu allocation metrics (#181) (c25eca9)

v1.26.20

1.26.20 (2025-05-13)

Bug Fixes

• properly set owner references for GPU resources using GVK scheme (#182) (c16750e)

v1.26.19

1.26.19 (2025-05-12)

Bug Fixes

• available field and owner reference issue in nodediscovery (#180) (bd64440)

v1.26.18

1.26.18 (2025-05-12)

Bug Fixes

• prevent release gpu failure (#179) (488c7e9)

v1.26.17

1.26.17 (2025-05-09)

Bug Fixes

• ensuring idempotency for the handlePodGPUCleanup function (#178) (6d484d8)

v1.26.16

1.26.16 (2025-05-08)

Bug Fixes

• calculate hash based on whole spec (#175) (aaa252d)

v1.26.15

1.26.15 (2025-05-07)

Bug Fixes

• recreate hypervisor pod after eviction (#173) (60d28e9)

v1.26.14

1.26.14 (2025-05-02)

Bug Fixes

• gpu info matching and node discovery issues (#169) (22e7d24)