]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
rework convert.py to read hyper-parameters from config.json (#1958)
authorErik Scholz <redacted>
Thu, 22 Jun 2023 12:20:47 +0000 (14:20 +0200)
committerGitHub <redacted>
Thu, 22 Jun 2023 12:20:47 +0000 (14:20 +0200)
commit7487137227eb32ed9b12156338b865cb29b2dfd1
treed65ed6238bf5a04519dc038114fe7cc332993720
parentbbca06e26949686d61a5126332680ba3cccf235c
rework convert.py to read hyper-parameters from config.json (#1958)

* Read hyper-parameters from HuggingFace-transformer config.json, if they exist, and fall back to guessing, like before otherwise.
  This allows converting open_llama 3B and other non-standard model designs.
convert.py