ziy/tpu-inference
TPU inference for vLLM, with unified JAX and PyTorch support.
Researcher and software developer in LLM, NLP and ML, currently at Google DeepMind.
TPU inference for vLLM, with unified JAX and PyTorch support.
Leveraging Procedural Knowledge for Task-oriented Search
A CSE based corpus preprocessing and indexing pipeline for TREC KBA task
A booklet to guide you how to (1) install and configure all the related softwares (Java, Eclipse, Git, Maven, etc.) from scratch, (2) understand UIMA concepts and learn how to implement a simple NLP tool based on UIMA, (3) build an information processing pipeline based on CSE
Java client for GoPubMed services used for BioASQ task.
Minimal Medline Baseline Indexer
A Jekyll theme with JavaScript powers.
Simple embedding based text classifier inspired by fastText, implemented in tensorflow
OAQA Biomedical Question Answering (BioASQ) System
Java-based API to the Semantic Knowledge Representation (SKR) Scheduler facility http://ii.nlm.nih.gov/Web_API/index.shtml
Personal repository for releases
UMLS Terminology Services API 2.0
Frequent Phrase Extractor from Lucene Index or Text Collection
Email and Calendar Agent
A PDTB-Styled End-to-End Discourse Parser
DisGeNET Analyzer
Minimal PubMed Central Baseline Indexer
Online pipeline for executing TREC KBA task
Evaluation for homework 2
Corpus proprocessing code, including analyzing corpus size, annotations, indexing, as well as prefetching related named entities for 29 given topics.
A "hello world" question answering pipeline for biomedical questions based on configuration space exploration framework
A Maven archetype buddles UIMA related dependencies, BioCreative GM task corpus, and some GM sample codes.