Skip to content

[bug fix] Compute fp16 batch norm in fp32 instead of downcasting the params #137

[bug fix] Compute fp16 batch norm in fp32 instead of downcasting the params

[bug fix] Compute fp16 batch norm in fp32 instead of downcasting the params #137

Triggered via pull request August 7, 2026 19:38
Status Success
Total duration 6m 36s
Artifacts

ci.yml

on: pull_request
Fit to window
Zoom out
Zoom in