https://github.com/daedalus/aiscripts

Scripts created with AI of white papers and publications
https://github.com/daedalus/aiscripts

ai bloomfilter cassowary-algorithm cuckoo-filter feistel-network gemm gpt karatsuba karatsuba-matrix-multiplication llm meta-gradient-descent noprop phylolm whitepaper zeedlm zeromerge

Last synced: 14 days ago
JSON representation

Scripts created with AI of white papers and publications

• Host: GitHub
• URL: https://github.com/daedalus/aiscripts
• Owner: daedalus
• Created: 2025-04-07T16:53:17.000Z (about 1 month ago)
• Default Branch: master
• Last Pushed: 2025-04-29T16:12:24.000Z (19 days ago)
• Last Synced: 2025-04-29T17:28:07.482Z (19 days ago)
• Topics: ai, bloomfilter, cassowary-algorithm, cuckoo-filter, feistel-network, gemm, gpt, karatsuba, karatsuba-matrix-multiplication, llm, meta-gradient-descent, noprop, phylolm, whitepaper, zeedlm, zeromerge
• Language: Jupyter Notebook
• Homepage:
• Size: 907 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# AIscripts #

**AIscripts** is a collection of experimental algorithms, machine learning components, and optimization tools inspired by various domains. This repository serves as a playground for exploring computational efficiency, mathematical models, and AI-driven solutions.

![AIscripts](https://img.shields.io/badge/Status-Experimental-orange)

---

## Features

### Core Components

- **Matrix Operations**:

  `GEMM_tiled.py` - Tiled implementation of General Matrix Multiply (GEMM).

  `KMM.py` - Karatsuba Matrix Multiplication utilities.

  `MGD.py` - Meta Gradient Descent optimizer.  

- **AI/ML Tools**:  

  `SeedLM.py` - Seed and coefficient selection algorithm for a single weight block.

  `aicommit.py` - AI-assisted Git commit message generator.

  `barcode_ssd_mobilenet_v1_dmp25_quant.tflite.runner.py` - TensorFlow Lite runner for SSD MobileNet-based barcode detection.  

  `ZeroMerge.py` - Parameter-Free KV Cache Compression for Memory-Efficient Long-Context LLMs.

  `NoProp.py` - Training Neural Networks without Back-propagation or Forward-propagation.

  `PhyloLM.ipynb` and `phylolm.py` - Inferring the Phylogeny of Large Language Models and Predicting their Performances in Benchmarks.

  `DLFloat.py` - Lossless LLM Compression for Efficient GPU Inference via Dynamic-Length Float

- **Probabilistic Data Structures**:

  `bloomfilter.py` - Classic Bloom filter implementation.

  `cuckoofilter.py` - Space-efficient Cuckoo filter.

  

- **Optimization & Cryptography**:   

  `minimal_feistel_network.py` - Compact Feistel network cipher.

  

- **Combinatorial Algorithms**:  

  `polyomino_tiling.py` - Polyomino tiling solver.  

  `cassowary.py` - Constraint-solving algorithm implementation.  

- **Misc**:

  `calcDa.py` - Density altitude calculator.

  `mdnscan.py` - mdns simple scanner.  

  `mbf.py` - Minimal bloom filter implementation. 

---

## Getting Started

### Prerequisites

- Python 3.8+

- `pip install -r requirements.txt`  

  *(Example dependencies: NumPy, TensorFlow Lite, PyTorch)*

### Basic Usage

```bash

# Run matrix multiplication benchmark

python GEMM_tiled.py --size 1024

# Generate AI-assisted commit message

python aicommit.py --diff