vllm/entrypoints/openai/chat_completion/serving.py · d084e9fca7d5d40cbb62eb5fe8ab5cbc6c769cf0 · OpenDAS / vllm_cscc

[MODEL] Fix handling of multiple channels for gpt-oss with speculative decoding (#26291) · d084e9fc

Aleksandr Samarin authored Jan 14, 2026


Signed-off-by: Aleksandr Samarin <astrlrd@nebius.com>
Signed-off-by: southfreebird <yvorott@gmail.com>
Co-authored-by: southfreebird <yvorott@gmail.com>

d084e9fc

serving.py 82.5 KB

Replace serving.py