• Serve a ML model on a single machine with Flask + Gunicorn vs. FastAPI + Uvicorn

  8 min read   ·   February 1, 2024

  2024   ·   dev     ·   flask,   gunicorn,   fastapi,   uvicorn  

• I find Ray a powerful framework for parallel computing

  10 min read   ·   January 1, 2024

  2024   ·   dev     ·   multiprocessing,   ray  

• Different techniques for optimizing LLM inference

  4 min read   ·   October 4, 2023

  2023   ·   dev     ·   NLP,   LLM   Inference  

• Cheat Sheet of NLP Practitioner

  137 min read   ·   September 8, 2023

  2023   ·   research     ·   NLP,   AI  

• Recap of The Very Large Database (VLDB) Conference 2023

  14 min read   ·   September 1, 2023

  2023   ·   research     ·   NLP,   Database  

• Table Representation Learning with Transformer

  10 min read   ·   August 5, 2023

  2023   ·   research     ·   NLP,   Table_Representation_Learning,   AI  

• Experimental benchmarks on the GPU requirements for the training/fine-tuning of LLMs.

  2 min read   ·   June 12, 2023

  2023   ·   dev     ·   language_model,   gpu  

• PoTM - Emergent World Representations - Exploring a Sequence Model Trained on a Synthetic Task

  5 min read   ·   May 1, 2023

  2023   ·   research     ·   paper_of_the_month,   language_model,   explainable_ai  

• PoTM - Understanding Dataset Difficulty with V-Usable Information

  8 min read   ·   December 6, 2022

  2022   ·   research     ·   paper_of_the_month,   information_theory,   machine_learning  

• What I've learned from finding ways to accelerate the inference of a Transformer model.

  13 min read   ·   November 10, 2022

  2022   ·   dev     ·   Optimization,   ONNX,   ONNX_Runtime,   Huggingface_Optimum,   Transformer.