]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 12b1750)
llama : recognize IBM Granite 3.3 FIM tokens (#12988)
authorMikko Juola <redacted>
Thu, 17 Apr 2025 08:37:05 +0000 (01:37 -0700)
committerGitHub <redacted>
Thu, 17 Apr 2025 08:37:05 +0000 (11:37 +0300)
commit971f245b3b5f3f55991bb779cb541b00f82eea1d
tree84e3900a7c15b1f925b38f78831bb30c42cafe2dtree
parent12b17501e6015ffe568ac54fdf08e6580833bf1bcommit | diff
llama : recognize IBM Granite 3.3 FIM tokens (#12988)

The Granite's FIM tokens are very similar to Qwen's; it's just that
they use underscore instead of a dash. So <fim_middle> for example
instead of <fim-middle>.

Opening up tokenizer_config.json in ibm-granite/granite-3.3-8b-base
shows:

```
    "<fim_prefix>",
    "<fim_middle>",
    "<fim_suffix>",
    "<fim_pad>",
    ...
    "<reponame>",
```
src/llama-vocab.cpp diff | blob | history
Packaging of ggml-org/llama.cpp