[V1][Spec Dec Bug Fix] Respect Spec Dec Method Specification#16636
[V1][Spec Dec Bug Fix] Respect Spec Dec Method Specification#16636vllm-bot merged 1 commit intovllm-project:mainfrom
Conversation
Signed-off-by: Bryan Lu <yuzhelu@amazon.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 🚀 |
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com>
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com> Signed-off-by: Yang Wang <elainewy@meta.com>
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com>
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com>
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com> Signed-off-by: Agata Dobrzyniewicz <adobrzyniewicz@habana.ai>
…oject#16636) Signed-off-by: Bryan Lu <yuzhelu@amazon.com> Signed-off-by: Mu Huai <tianbowen.tbw@antgroup.com>
Currently, we try to automatically detect the speculative decoding method. For EAGLE, it detects via the draft model checkpoint folder. However, if the folder name does not have the
eagle-substring, the current logic will make the methoddraft_modeleven if one has directly specify the method aseagle, which leads to subsequent errors. This fix avoids this confusion.@LiuXiaoxuanPKU