astellon

A Bayesian Permutation training deep representation learning method for speech enhancement with variational autoencoder	http://arxiv.org/abs/2201.09875v1
A COMPARISON OF DISCRETE AND SOFT SPEECH UNITS FOR IMPROVED VOICE CONVERSION	http://arxiv.org/abs/2111.02392v1
A Configurable Multilingual Model is All You Need to Recognize All Languages	http://arxiv.org/abs/2107.05876v1
A Framework for Private Communication with Secret Block Structure	http://arxiv.org/abs/2110.04345v1
A GENERAL FRAMEWORK FOR DISTRIBUTED INFERENCE WITH UNCERTAIN MODELS	http://arxiv.org/abs/2011.10669v1
A GENERALIZED HIERARCHICAL NONNEGATIVE TENSOR DECOMPOSITION	http://arxiv.org/abs/2109.14820v2
A likelihood ratio based domain adaptation method for E2E models	http://arxiv.org/abs/2201.03655v1
A Maximal Correlation Approach to Imposing Fairness in Machine Learning	http://arxiv.org/abs/2012.15259v1
A METHOD TO REVEAL SPEAKER IDENTITY IN DISTRIBUTED ASR TRAINING, AND HOW TO COUNTER IT	http://arxiv.org/abs/2104.07815v1
A NEW DATA AUGMENTATION ME

astellon

LibriSpeechのWebDataset化

まとめ

• Librispeechという約60GBぐらいの音声データセットを
• webdatasetという形式に変換して
• Disk I/Oを効率化する

アバウトな測定結果

astellon

FROM nvcr.io/nvidia/cuda:11.6.2-cudnn8-devel-ubuntu20.04

RUN apt-get update && \
    apt-get -y install python3 python3-pip && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

RUN python3 -m pip install jax[cuda] flax -f https://storage.googleapis.com/jax-releases/jax_releases.html
RUN python3 -m pip install tensorflow tensorflow_datasets wandb hydra-core

astellon

FROM nvcr.io/nvidia/cuda:11.6.2-runtime-ubuntu20.04

RUN apt-get update && \
    apt-get -y install python3 python3-pip && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

RUN python3 -m pip install jax[cuda] flax -f https://storage.googleapis.com/jax-releases/jax_releases.html

ENTRYPOINT [ "python3" ]

astellon

FROM node:current-alpine3.15

RUN apk --no-cache add git
RUN git clone https://github.com/lojban/ilmentufa.git

WORKDIR /ilmentufa

RUN rm package-lock.json

RUN npm install

astellon

# generate fractional Brownian motion (fBM)
# reference: https://thebookofshaders.com/13 by Patricio Gonzalez Vivo

using LinearAlgebra
using Images, FileIO

function fract(x)
    x - floor(x)
end

astellon

# Julia implementation of [rVAD](https://www.sciencedirect.com/science/article/abs/pii/S0885230819300920).
# This implementation is based on [the official Python implementation](https://github.com/zhenghuatan/rVAD/blob/master/rVADfast_py_2.0/rVAD_fast.py)
# and the original license is [here](https://github.com/zhenghuatan/rVAD/blob/master/rVADfast_py_2.0/LICENSE).

using Base.Iterators: partition
using DSP, WAV

function loadwav(path)
  xs, fs = wavread(path)
  xs = reshape(sum(xs, dims=2)./size(xs, 2), :)  # make mono

astellon

FROM kalilinux/kali

ARG OVPN_CONF
ENV DEBIAN_FRONTEND noninteractive

RUN apt-get update && apt-get install kali-linux-default -yqq -o Dpkg::Options::="--force-confdef" -o Dpkg::Options::="--force-confold"

COPY ./entrypoint.sh /
COPY ${OVPN_CONF} /

astellon

Commodore 64 開発ツールなど

GitHub上にあって最近でもメンテナンスされてそうなもの中心。Linuxで動くことが最優先。

Vice

エミュレータ。基本みんなこれ。

dasm

astellon

def cycle(iterable):
  it = iter(iterable)
  while True:
    try:
      yield next(it)
    except StopIteration:
      it = iter(iterable)
      continue