Skip to content
hpu-gaudi2

Enabled high-performance Automatic Tensor Parallelism (auto TP) for the MoE models on multiple GPUs/HPUs #1728

Sign in to view logs

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 #1728

This workflow is awaiting approval from a maintainer in #6964
Triggered via pull request February 27, 2025 08:32
@gyou2021gyou2021
synchronize #6964
gyou2021:autoTP_Qwen2Moe_DeepSeekv2
Status Action required
Total duration –
Artifacts –
This workflow is awaiting approval from a maintainer in #6964

hpu-gaudi2.yml

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