HN Explorer

bornaahz

↗ HN profile [ 45.8 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. 1 points 0
  2. 1 points 0
  3. 1 points 0
  4. 1 points 0
  5. 1 points 0
  6. 1 points 0
  7. Show HN: Flaim – pre-trained vision backbones for Flax (github.com)

    flaim is a library of state-of-the-art pre-trained vision models, plus common deep learning modules in computer vision, for JAX/Flax. It exposes a host of diverse image models through a straightforward interface with an…

    3 points 0
  8. Show HN: OpenCLIP Jax – Implementation of OpenAI's Clip in Jax/Flax (github.com)

    open_clip_jax is an open source JAX/Flax implementation of OpenAI's CLIP, including image and text towers, pre-trained parameters, training utilities that can scale to TPU Pods, and more. It is inspired by but not…

    1 points 0
  9. Show HN: Minbpe-hs – byte-level byte pair encoding (BPE) in Haskell (github.com)

    minbpe-hs is a port of Andrej Karpathy's concise byte-level byte pair encoding (BPE) implementation, minbpe, to Haskell. During training, BPE compresses a sequence by repeatedly finding the most frequent pair of…

    2 points 0
  10. Show HN: Attorch – PyTorch's nn module written in Python using OpenAI's Triton (github.com)

    attorch is a subset of PyTorch's nn module, written purely in Python using OpenAI's Triton. Its goal is to be an easily hackable, self-contained, and readable collection of neural network modules whilst maintaining or…

    4 points 0