https://github.com/jrhea/anomlee

ANOMLEE: A Neat-O ML EE (pronounced a·nom·a·ly)
https://github.com/jrhea/anomlee

c ethereum machine-learning python rust wasm

Last synced: 21 days ago
JSON representation

ANOMLEE: A Neat-O ML EE (pronounced a·nom·a·ly)

• Host: GitHub
• URL: https://github.com/jrhea/anomlee
• Owner: jrhea
• License: apache-2.0
• Created: 2019-12-06T02:22:08.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2022-12-08T03:19:18.000Z (almost 2 years ago)
• Last Synced: 2024-10-07T20:05:09.541Z (about 1 month ago)
• Topics: c, ethereum, machine-learning, python, rust, wasm
• Language: Jupyter Notebook
• Homepage: anomlee.vercel.app
• Size: 66.4 KB
• Stars: 8
• Watchers: 3
• Forks: 1
• Open Issues: 8
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
## ANOMLEE: A Neat-O Machine Learning Execution Environment

:muscle: by Quilt's [ewasm-rt](https://github.com/quilt/ewasm-rt)

This repo demonstrates how a trained ML model could run inside an Ethereum 2 EE.

## Prerequisits

Install LLVM

```bash

brew install llvm

echo 'export PATH="/usr/local/opt/llvm/bin:$PATH"' >> ~/.bash_profile

```

Install the WebAssembly Binary Toolkit

```bash

brew install wabt

```

Install Python Environment & Dependencies

```

python3 -m venv env

source ./env/bin/activate

pip3 install -r requirements.txt

```

## Build 

```bash

make all

```

## Random Forest Classifier (iris model)

This model that is trained to determine which of the following types of iris flowers:

- setosa

- versicolor

- virginica

 based on the following design variables:

- sepal length (cm)	

- sepal width (cm)	

- petal length (cm)	

- petal width (cm)

**Benchmarks:**

- **Python:** 103307 microseconds

- **eWasm:** 551.849 microseconds

- **C:** 1 microsecond

**Binary size:**

- 826 bytes

## Handwritten Digit Classifier (digit model)

This model uses a sequential neural network model to classify handwritten digits from the MINST database

> TODO: finish

## Run Benchmarks (all models)

```bash

$ make benchmark

########## Python Benchmark: ###########

Execution time: 103307 microseconds.

 0    setosa

Name: species, dtype: category

Categories (3, object): [setosa, versicolor, virginica] 

########## eWasm Benchmark: ###########

    Finished release [optimized] target(s) in 0.03s

     Running target/release/deps/anomlee-0de6000a8c14c88e

running 1 test

Execution Time: 551.849µs

test tests::test ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out

########## C Benchmark: ##########

Execution time: 1 microseconds

Probabilities: 

1.000000 0.000000 0.000000 

Model Predicts: 

setosa    

```