Pinned Loading

1. infer-lab infer-lab Public

   LLM inference engine internals: PagedAttention, RadixAttention, continuous batching, speculative decoding - kernels to distributed systems

   Python

2. turbovec turbovec Public

   Fast, from-scratch ANN vector search engine — SIMD-accelerated C++ (HNSW) core with a Python-first API. 100% recall at ~11k QPS.

   Python

3. LoomIndex LoomIndex Public

   A memory-efficient C++20 web crawler designed for extreme scalability. Implements probabilistic URL deduplication (Bloom Filter) to minimize RAM footprint and features a non-blocking asynchronous I…

   C++ 1

4. OmniSeg-Audio-Pipeline OmniSeg-Audio-Pipeline Public

   High-performance multimodal pipeline synchronizing Meta's SAM 2 (Vision) and MIT's AST (Audio). Features O(1) resource management, temporal video slicing, and automated JSON metadata orchestration …

   Python

5. Fast-Prototyping-of-a-GenAI-App-with-Streamlit Fast-Prototyping-of-a-GenAI-App-with-Streamlit Public

   🤖 Streamlit-based GenAI application featuring RAG and Snowflake integration for advanced data ingestion, cleaning, and AI-driven insights.

   Python 1