PR #1473 whisper : support ggml_conv with CUDA and Metal

whisper : support ggml_conv with CUDA and Metal #1473

ggerganov merged 5 commits into ggml-backend-no-sched from ggml-conv

ggml : add CUDA support for ggml_conv

81506268

whisper : remove ggml_repeat for conv bias + single backend

000b952c

ggerganov commented on 2023-11-10

cuda : fix im2col kernel

9c1ddc77

metal : add im2col support + mul mat-vec f16 x f16

997f7cbf

bench-all : add q4 models

7a91a3ba

ggerganov marked this pull request as ready for review 2 years ago

ggerganov changed the title ~~whisper : support ggml_conv with CUDA~~ whisper : support ggml_conv with CUDA and Metal 2 years ago

ggerganov merged 933c5bef into ggml-backend-no-sched 2 years ago

Reviewers

slaren

FSSRepo

Assignees

No one assigned

Labels

None yet

Milestone

No milestone