llama.cpp
ggml : fix padding in timestep embedding kernels
#15932

Merged

ggml : fix padding in timestep embedding kernels #15932

danbev merged 7 commits into ggml-org:master from danbev:ggml-timestep-embedding-backends

danbev requested a review from

0cc4m 9 days ago

github-actions added Nvidia GPU

github-actions added Vulkan

github-actions added ggml

github-actions added SYCL

github-actions added Apple Metal

github-actions added OpenCL

ggerganov approved these changes on 2025-09-11

danbev force pushed from 004eaba8 to d304f6bf 4 days ago

ggml : remove adding extra dim timestep embedding

c40a17a5

ggml-cuda : fix padding in timestep embedding kernel

5db73a8b

ggml-metal : fix padding in timestep embedding kernel

b55c1943

ggml-opencl : fix padding in timestep embedding kernel

a1de8bd8

ggml-sycl : fix padding in timestep embedding kernel

7f124412

ggml-vulkan : fix padding in timestep embedding kernel

cb6b133e

ggml-cpu : fix padding in timestep embedding function

aa2c30ca

danbev force pushed from d304f6bf to aa2c30ca 4 days ago

danbev removed review request from

0cc4m 4 days ago

danbev requested a review from

ggerganov 4 days ago

ggerganov approved these changes on 2025-09-16

danbev merged 3913f873 into master 4 days ago

Reviewers

ggerganov

Assignees

No one assigned

Labels

Nvidia GPU Vulkan ggml SYCL Apple Metal OpenCL

Milestone

No milestone