]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
llama : support Llama 3 HF conversion (#6745)
authorPedro Cuenca <redacted>
Sun, 21 Apr 2024 11:50:41 +0000 (13:50 +0200)
committerGitHub <redacted>
Sun, 21 Apr 2024 11:50:41 +0000 (14:50 +0300)
commitb97bc3966e852adb626c90be64fd48282800f504
tree178656d15821205889fa03ec603c7327facbb265
parentb8109bc0139f15a5b321909f47510b89dca47ffc
llama : support Llama 3 HF conversion (#6745)

* Support Llama 3 conversion

The tokenizer is BPE.

* style

* Accept suggestion

Co-authored-by: Sourab Mangrulkar <redacted>
* llama : add llama_token_is_eog()

ggml-ci

* llama : auto-detect more EOT tokens when missing in KV data

* convert : replacing EOS token is a hack

* llama : fix codegemma EOT token + add TODOs

* llama : fix model type string for 8B model

---------

Co-authored-by: Sourab Mangrulkar <redacted>
Co-authored-by: Georgi Gerganov <redacted>
20 files changed:
convert-hf-to-gguf.py
convert.py
examples/batched.swift/Sources/main.swift
examples/batched/batched.cpp
examples/beam-search/beam-search.cpp
examples/infill/infill.cpp
examples/llama.android/app/src/main/cpp/llama-android.cpp
examples/llama.swiftui/llama.cpp.swift/LibLlama.swift
examples/llava/llava-cli.cpp
examples/lookahead/lookahead.cpp
examples/lookup/lookup.cpp
examples/main/main.cpp
examples/parallel/parallel.cpp
examples/passkey/passkey.cpp
examples/server/server.cpp
examples/server/utils.hpp
examples/simple/simple.cpp
examples/speculative/speculative.cpp
llama.cpp
llama.h