]> git.djapps.eu Git - pkg/ggml/sources/ggml/commit
ggml : use WARP_SIZE/2 for argmax reduction offset (llama/18092)
authorAadeshveer Singh <redacted>
Wed, 17 Dec 2025 03:47:01 +0000 (09:17 +0530)
committerGeorgi Gerganov <redacted>
Wed, 17 Dec 2025 11:55:04 +0000 (13:55 +0200)
commit23c6c8ae9f0243b4e25a666ace24c537dfb6ca0e
treec41193ec6a7a680ccba25c20902a8b1761f4d501
parentb821be8e3e963e149c2aee9b04292efed9765535
ggml : use WARP_SIZE/2 for argmax reduction offset (llama/18092)
src/ggml-cuda/argmax.cu