牛首山

Popular repositories Loading

1. FormalASR FormalASR Public

   An end-to-end ASR model, transcribing spoken Chinese to formal text.

   Python 22

2. PS4 PS4 Public

   Proxy-Supervised Joint Training for Real Target Speaker Extraction

   Python 11 3

3. homie homie Public

   Homie — 自进化智能助手 CLI

   Python 6

4. DAVE DAVE Public

   DAVE: audio-visual speech enhancement & target speaker extraction for real-world meetings — audio-only TIGER-M backbone (2.56M) + four-cue speaker attribution. ISCSLP 2026 AVSE Challenge: Track1 4t…

   Python 6 2

5. wecom-tools-dify-plugin wecom-tools-dify-plugin Public

   企微dify插件

6. dsh-llm-as-a-verifier dsh-llm-as-a-verifier Public

   LLM-as-a-Verifier for DeepSeek Harness: fine-grained reward, Probabilistic Pivot Tournament best-of-N selection, and per-step progress tracking as agent tools.

   TypeScript