]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
CUDA: faster softmax via shared memory + fp16 math (llama/4742)
authorJohannes Gäßler <redacted>
Tue, 9 Jan 2024 07:58:55 +0000 (08:58 +0100)
committerGeorgi Gerganov <redacted>
Thu, 11 Jan 2024 19:45:42 +0000 (21:45 +0200)
commitcdaa17ad0cc96dd1b4b782cff7d96bf1aa5c1fb0
treed14a0480ba75a759f47f484fa8f5a2e02756c195
parent979cc23b345006504cfc1f67c0fdf627805e3319
CUDA: faster softmax via shared memory + fp16 math (llama/4742)
src/ggml-cuda.cu
tests/test-backend-ops.cpp