CUDA: don't convert BF16 weights to FP32#1174

Merged

JohannesGaessler merged 4 commits intoggml-org:masterfrom

CISC:cuda-bf16-support

Apr 4, 2025

Commits on Apr 2, 2025

add bf16 support
CISC
authored

Commits on Apr 3, 2025

Commits on Apr 4, 2025

move functionality into convert_unary with constexpr
CISC
authored