You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Implementation of Flash-DLM (paper: FlashDLM: Accelerating Diffusion Language Models via Efficient KV Caching and Guided Diffusion). Provides training-free methods to accelerate diffusion language model inference.
A Python tool that extracts telemetry data from various space launch videos using computer vision and OCR. Analyzes flight parameters (speed, altitude, engine status), calculates derived metrics, and generates visualizations for comparing launch performance characteristics. Features parallel processing and a user-friendly command-line interface.
Accelerating LLM inference with techniques like speculative decoding, quantization, and kernel fusion, focusing on implementing state-of-the-art research papers.