Developer Profile

A.V.

@slckl

Latvia

GitHub

Impact Score Overview

Developer Impact & Statistics

Final Score

41 / 100Raw: 255.00

Repository Score

19 / 100Raw: 23.00

PR Score

64 / 100Raw: 533.00

Contribution Score

34 / 100Raw: 51.00

Score Weight Distribution

finalScore = repoScore * 0.45 + prScore * 0.45 + contributionScore * 0.10

Repository Score (45%)4% of final signal

Pull Request Score (45%)94% of final signal

Contribution Score (10%)2% of final signal

Top Work

Top Work & Highlights

Top Repositories

Most impactful repositories and pull requests

#1gstreamed_rust_inference

stars10

forks0

Rust 10000%

Score

#2candle_rf_detr

stars1

forks0

Python 7100%Rust 2900%

Score

#3candle_xps

stars1

forks0

Python 6700%Rust 3300%

Score

Top Pull Requests

Most impactful repositories and pull requests

#1CPU Conv2d: separate module, tiled im2col, specialization

in huggingface/candle

stars on the PR repository21013

+435/-129

Rust 7600%Cuda 900%Metal 600%

109

Score

#2[RNNLM] old iteration model cleanup

in kaldi-asr/kaldi

stars on the PR repository15475

+182/-11

Shell 4700%C++ 3700%Python 900%

Score

#3Sort on cuda fails when tensor size exceeds 1024

in huggingface/candle

stars on the PR repository21013

+48/-26

Rust 7600%Cuda 900%Metal 600%

Score

Community Contributions

comments

#1Issue

Conv2d cpu performance is somewhat slow

huggingface/candle

stars21013

comments18

Score

#2Issue

#[serde(flatten)] doesn't work with #[serde(default)] for externally tagged enums

serde-rs/serde

stars10809

comments9

Score

#3Issue

Add baseline visualizations

jettify/pytorch-optimizer

stars3170

comments5

Score

Scoring Transparency

▼

Repositories analyzed

Pull requests analyzed

Merged external PRs counted

Own-repo PRs ignored

Unmerged PRs ignored

Unique external repositories

Issues analyzed

External issues counted

Discussions analyzed

External discussions counted

How this score was calculated

View the full scoring methodology with formulas, penalties, and step-by-step flow.

View full methodology