]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: a686171)
llama : add support for GPT2, Bloom and CodeShell tied word embeddings (#12456)
authorSigbjørn Skjæret <redacted>
Wed, 19 Mar 2025 08:08:49 +0000 (09:08 +0100)
committerGitHub <redacted>
Wed, 19 Mar 2025 08:08:49 +0000 (09:08 +0100)
commit108e53c2f1b57b79c8bcf9773fba5a82e2f1eeb5
treeeeb18f936d00c7935258725692fb878266f00056tree
parenta686171ea71ed8cb8a324850d146cb65a001e141commit | diff
llama : add support for GPT2, Bloom and CodeShell tied word embeddings (#12456)

* Add support for GPT2, Bloom and CodeShell tied word embeddings

* Deduplicate tied word embeddings weights

* Workaround for incorrect weight map

It appears transformer.wte.weight is in the weight map even though the weights are not there, remove it if output weights are encountered first.

* check++

* fatfingers--
convert_hf_to_gguf.py diff | blob | history
src/llama-model.cpp diff | blob | history
Packaging of ggml-org/llama.cpp