If you believe there's a paper, problem, or project, especially in the domains of AI, CTF, or Competitive Programming that could benefit from an open-source implementation, I'd love to hear from you! Please use this formlink to share your ideas or reach out to me directly via mail or Discord.

• self-compressing-nn-jax JAX Implementation of the paperhttps://arxiv.org/pdf/2301.13142 ported from gehots tinygrad version.repo
• llama.jl Inference Llama 2 in one file of pure C. Nahh wait, now fresh in Julia!repo
• CoPE-cuda Contextual Position Encoding but with some custom CUDA Kernelshttps://arxiv.org/abs/2405.18719repo
• Competitive-Coding-Benchmark Benchmark LLMs against a set of high quality representative competitive coding problems sourced from cses.firepo
• nngdb An terminal based neural-network Interpretability-tool that feels like gdbrepo
• cudatiger An accelerated implementation of the Tiger optimizer for PyTorch, supercharged with Triton for enhanced CUDA GPU efficiency. More Memory efficient than Adam.https://kexue.fm/archives/9512repo
• tinyfep A simplified and abstracted rendition of FEP+ (Free Energy Perturbation) for molecular simulations.https://www.schrodinger.com/products/feprepo
• betterwords.lol Expand human vocabulary with gpt-3.5-turbo or finetuned llama-2 in Reactlink

• juvi21/Kumiho-V1-rp-UWU-8B-GGUF An intelligent roleplay model finetuned mostly on llama-3.1 paper inspired synthetic claude-3.5-sonnet data.
• juvi21/Hermes-2.5-Yi-1.5-9B-Chat with GGUF and Quantization. Outperforms many models in it's class on AGIEval. More details in the model-card.
• juvi21/Hermes-2-Pro-Mistral-7B-infinity with GGUF and Quantization. More details in the model-card.

• cses-fi-competitive-coding-problems Highly representative set of competitive coding problems from cses.fi in a nice dataformat.

PinnedLoading

1. llama2.jlllama2.jlPublic

   Forked fromkarpathy/llama2.c

   Inference Llama 2 in one file of pure C. Nahh wait, now fresh in Julia!

   Python 23 1

2. CoPE-cudaCoPE-cudaPublic

   Contextual Position Encoding but with some custom CUDA Kernels https://arxiv.org/abs/2405.18719

   Python 22

3. PrimeIntellect-ai/toplocPrimeIntellect-ai/toplocPublic

   TOPLOC: is a novel method for verifiable inference that enables users to verify that LLM providers are using the correct model configurations and settings

   Python 17 5

4. init.viminit.vim

   1

   callplug#begin()

   2

   3

   " UI and Themes

   4

   Plug'drewtempelmeyer/palenight.vim'

   5

   Plug'morhetz/gruvbox'

5. cudatigercudatigerPublic

   An accelerated implementation of the Tiger optimizer for PyTorch, supercharged with Triton for enhanced CUDA GPU efficiency. More Memory efficient than Adam. https://kexue.fm/archives/9512

   Python 2

6. Competitive-Coding-BenchmarkCompetitive-Coding-BenchmarkPublic

   Benchmark LLMs against a set of high quality representative competitive coding problems sourced from cses.fi

   Python 2