CLASP: Contrastive Language-Speech Pretraining for Multilingual Multimodal Information Retrieval
-
Updated
Jun 27, 2025 - Jupyter Notebook
CLASP: Contrastive Language-Speech Pretraining for Multilingual Multimodal Information Retrieval
Source code of our paper "INSPIRE: A Benchmark for Instruction-Aware Speech Retrieval", Interspeech 2026 Long Paper
面向低资源方言开放词表检索的跨模态对齐与路由陷阱实验室:组合式零样本 OOV、语速不变跨模态注意力、类条件域对齐与 FSQ 离散晶格。Modular cross-modal dialect alignment benchmark with a live white-box Lab.
免训练方言关键词检索与白盒审计台:80ms/40ms 双网格防泄漏、Top-20 语境重排、Prefix KV-Cache 加速与五折阴性实验公开。Training-free Audio-LLM keyword spotting with a live white-box Audit Studio.
PostgreSQL database module for a Speech Retrieval System featuring relational schema design, SQL queries, indexing, and backend data management.
To associate your repository with the speech-retrieval topic, visit your repo's landing page and select "manage topics."