]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 2e89f76)
context : round n_tokens to next multiple of n_seqs when reserving (#14140)
authorcompilade <redacted>
Thu, 12 Jun 2025 06:56:04 +0000 (02:56 -0400)
committerGitHub <redacted>
Thu, 12 Jun 2025 06:56:04 +0000 (02:56 -0400)
commita20b2b05bce6622c585459ebf46f142f113d021c
tree066be3659fdf86d80fc7ccce485aeabeed8efde4tree
parent2e89f76b7af2c0b827be785e445f2e2b3e52e1cacommit | diff
context : round n_tokens to next multiple of n_seqs when reserving (#14140)

This fixes RWKV inference which otherwise failed
when the worst case ubatch.n_seq_tokens rounded to 0.
src/llama-context.cpp diff | blob | history
Packaging of ggml-org/llama.cpp