HardikVala

↗ HN profile [ 66.1 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. hey hn, wrote a python library for myself to visualize tensors. Makes learning and debugging deep learning code so much easier. Works seamlessly with colab/jupyter notebooks, and other python contexts. It's built on top…

  2. A minimal implementation of ModernBERT in pure C, inspired by karpathy's llama2.c. The core (tokenizer + inference code) is around 1000 lines of code with no dependencies (except OpenBLAS for fast matrix multiplication,…

  3. I constantly find myself having to print raw tensors (either by hand or dumping them to stdout), especially when reading pytorch / jax code, to understand the transformations, e.g. for something like, `x =…

  4. Considering the server version of llama.cpp for a commercial use case over bulkier options like vllm. But wondering if it's been battle-tested in production environments.

  5. A variant of this question seems to get asked every 6 mo. but so far, I haven't seen this question tackled directly: If I want to learn the concepts and fundamentals of AI from first principles, what educational…