[ICASSP 2022] Improving End-to-End Contextual Speech Recognition with Fine-Grained Contextual Knowledge Selection
-
Updated
Jul 14, 2026
[ICASSP 2022] Improving End-to-End Contextual Speech Recognition with Fine-Grained Contextual Knowledge Selection
OmniFlow — A Unified Framework for Training-Free Omni-Modal LLM Orchestration. Three pillars (LLM orchestration, ThinkOmni decoding, Foresight streaming) compose into one training-free, model-agnostic stack. 69/69 tests pass offline. FastAPI dashboard + 7-expert pool.
OmniFlow: A Unified Framework for Training-Free Omni-Modal LLM Orchestration combining LLM Orchestration, ThinkOmni Collaborative Decoding, and Foresight Streaming Planning
To associate your repository with the collaborative-decoding topic, visit your repo's landing page and select "manage topics."