Skip to content

[Hetero] Support heterogeneous training strategy with different `dp_s… #681

[Hetero] Support heterogeneous training strategy with different `dp_s…

[Hetero] Support heterogeneous training strategy with different `dp_s… #681

Triggered via push September 24, 2024 03:03
Status Cancelled
Total duration 1h 50m 7s
Artifacts

test.yml

on: push
Matrix: flagscale-unit-tests
Matrix: megatron-unit-tests
Matrix: functional-tests-train
Matrix: functional-tests-hetero
Waiting for pending jobs
flagscale-coverage-test  /  test-coverage
flagscale-coverage-test / test-coverage
megatron-coverage-test  /  test-coverage
megatron-coverage-test / test-coverage
Fit to window
Zoom out
Zoom in

Annotations

3 errors
train-mixtral / functional-test
Process completed with exit code 1.
train-aquila / functional-test
FailFast: cancelling since parallel instance has failed
train-aquila / functional-test
The operation was canceled.