[Doc] Updated Llama section in tool calling docs to have llama 3.2 config info #16857
Conversation
Signed-off-by: jmho <jaylenho734@gmail.com>
|
👋 Hi! Thank you for contributing to the vLLM project. 💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels. Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can either: Add 🚀 |
Signed-off-by: jmho <jaylenho734@gmail.com>
github-actions
bot
added
the
ready
…nfig info (vllm-project#16857) Signed-off-by: jmho <jaylenho734@gmail.com> Signed-off-by: Yang Wang <elainewy@meta.com>
…nfig info (vllm-project#16857) Signed-off-by: jmho <jaylenho734@gmail.com>
…nfig info (vllm-project#16857) Signed-off-by: jmho <jaylenho734@gmail.com>
…nfig info (vllm-project#16857) Signed-off-by: jmho <jaylenho734@gmail.com> Signed-off-by: Agata Dobrzyniewicz <adobrzyniewicz@habana.ai>
…nfig info (vllm-project#16857) Signed-off-by: jmho <jaylenho734@gmail.com> Signed-off-by: Mu Huai <tianbowen.tbw@antgroup.com>
2 participants
The current tool calling docs for the Llama models notes that only the Llama 3.1 models are supported for JSON based tool calling even though there is support for the Llama 3.2 models as well.
This PR removes a mention of a chat template that no longer exists, adds the llama 3.2 models under supported models, and provides additional info about the differences between the 3.1 and 3.2 json chat templates.