Skip to content

split ReplicatedLinear used in MLA prefill computing along hidden_states[0] to save duplicated computing on all devices #9718

split ReplicatedLinear used in MLA prefill computing along hidden_states[0] to save duplicated computing on all devices

split ReplicatedLinear used in MLA prefill computing along hidden_states[0] to save duplicated computing on all devices #9718

This workflow is awaiting approval from a maintainer in #3688
Triggered via pull request February 25, 2025 04:55
Status Action required
Total duration
Artifacts
This workflow is awaiting approval from a maintainer in #3688

lint.yml

on: pull_request
lint
lint
Fit to window
Zoom out
Zoom in