kubeflow / trainer / 15978264926

30%

master: 51%

push

github

[release-2.0] fix(plugins): Fix some errors in torchtune mutation process. (#2693)

* fix(plugins): fix errors in trainer command mutation of torchtune.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(plugins): remove config file format suffix.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(test): update UTs.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(initializer): Update the workspace of dataset/model initializer.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(plugins): parse nproc_per_node from GPU resource.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(torchtune): Add bitsandbytes dependency in requirements.txt

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(lint): fix lint error.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* fix(torchtune): Remove unnecessary num_proc_per_node calculation.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

* test(torch): Update invalid parameters.

Signed-off-by: Electronic-Waste <2690692950@qq.com>

---------

Signed-off-by: Electronic-Waste <2690692950@qq.com>
Co-authored-by: Electronic-Waste <2690692950@qq.com>

18 of 18 new or added lines in 1 file covered. (100.0%)

901 of 3077 relevant lines covered (29.28%)

0.35 hits per line