Skip to content

Enabled high-performance Automatic Tensor Parallelism (auto TP) for the MoE models on multiple GPUs/HPUs #1742

Enabled high-performance Automatic Tensor Parallelism (auto TP) for the MoE models on multiple GPUs/HPUs

Enabled high-performance Automatic Tensor Parallelism (auto TP) for the MoE models on multiple GPUs/HPUs #1742

Re-run triggered February 28, 2025 03:37
Status Failure
Total duration 3m 3s
Artifacts

hpu-gaudi2.yml

on: pull_request
unit-tests
0s
unit-tests
Fit to window
Zoom out
Zoom in

Annotations

2 errors
hpu-gaudi2
Internal server error. Correlation ID: e3af4fc2-ee60-44b6-97ee-da05cdd005c9
unit-tests
The job was not acquired by Runner of type self-hosted even after multiple attempts