llama.cpp
ggml webgpu: actually add softmax, fix rms_norm offset
#16400

Merged

ggml webgpu: actually add softmax, fix rms_norm offset #16400

reeselevine merged 3 commits into ggml-org:master from reeselevine:master

implement soft_max

6464269f

github-actions added ggml

CISC approved these changes on 2025-10-03

Fix soft_max data race

ee47adb3

Temporary fix, wait on each submit

84770b49

reeselevine merged 35266573 into master 6 days ago

Reviewers

CISC

Assignees

No one assigned

Labels

ggml

Milestone

No milestone