Stop the generation when <|eom_id|> token is encountered (needed for llama 3.1 tool call support) #8858
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ggerganov
approved these changes
Aug 5, 2024
4 tasks
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this pull request
Aug 7, 2024
… Llama 3.1 tool call support (ggerganov#8858) * gguf-py, llama : add constants and methods related to Llama-3.1 <|eom_id|> token * llama : find Llama-3.1 <|eom_id|> token id during vocab loading * llama-vocab : add Llama-3.1 <|eom_id|> token to the set of tokens stopping the generation --------- Co-authored-by: Stanisław Szymczyk <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds support for <|eom_id|> token introduced by llama 3.1 models. It adds the EOM token to the list of tokens that stop the generation. This is necessary to allow proper tool call support, see https://llama.meta.com/docs/model-cards-and-prompt-formats/llama3_1/ for more details.
Note that it doesn't add any tool call support in llama.cpp, it only stops the generation after <|eom_id|> to allow implementation of tool calls in other software that uses llama.cpp for inference.
I don't feel confident enough to tinker with
LlamaModel::set_vocab()inconvert_hf_to_gguf.pyscript to explicitly set the EOM token value during conversion, so it's currently found during vocabulary loading like EOT tokens.I created a simple script allowing to test the llama 3.1 tool calling with llama-server: https://github.com/fairydreaming/tlcl