AI Rank
EN
← 开源
seth814

Audio-Classification

Code for YouTube series: Deep Learning for Audio Classification

TutorialsML/AI fundamentalsJupyter Notebook
在 GitHub 打开
增长势头
+024 小时新增 Star0.0%
586
Star
191
Fork
+0
本周
2
贡献者
创建于 2018-10-20 · 更新于 2026-08-30 · 今日第 16540 名
主要开发者
seth814

seth814/Audio-Classification 排名第 16,540，有 586 个星标。

同类别比较

seth814/Audio-Classification 排名第 16,540，此处有 8 个同类条目；avidevelops/claude-architect-exam-prep 排名第 16,533。

同类别比较
名称排名指标
jhavl/dkt16,471
语言:
Jupyter Notebook
星标数:
589
分叉数:
69
简介:
A Tutorial on Manipulator Differential Kinematics
davidrosenberg/mlcourse16,506
语言:
Jupyter Notebook
星标数:
587
分叉数:
266
简介:
Machine learning course materials.
apeman/awesome_computer_science16,532
星标数:
586
分叉数:
52
简介:
The complete syllabus of Computer Science and Engineering. Roadmap, Checklist for Beginners.
avidevelops/claude-architect-exam-prep16,533
星标数:
586
分叉数:
131
floodsung/DRL-FlappyBird16,548
语言:
Python
星标数:
585
分叉数:
321
简介:
Playing Flappy Bird Using Deep Reinforcement Learning (Based on Deep Q Learning DQN using Tensorflow)
gaoisbest/NLP-Projects16,549
语言:
OpenEdge ABL
星标数:
585
分叉数:
152
简介:
word2vec, sentence2vec, machine reading comprehension, dialog system, text classification, pretrained language model (i.e., XLNet, BERT, ELMo, GPT), sequence labeling, information retrieval, information extraction (i.e., entity, relation and event extraction), knowledge graph, text generation, network embedding
huxiaoman7/PaddlePaddle_code16,550
语言:
Python
星标数:
585
分叉数:
202
简介:
用PaddlePaddle和Tensorflow实现常用的深度学习算法
Jcharis/Machine-Learning-Web-Apps16,563
语言:
Jupyter Notebook
星标数:
584
分叉数:
562
简介:
Building and Embedding Machine Learning Model into a Web App(With Flask,Streamlit,etc)
关键参数与规格
关键参数与规格
参数值
语言Jupyter Notebook
类别Tutorials
子类别ML/AI fundamentals
关键词Audio production
星标数586
分叉数191
贡献者2
创建日期2018-10-20
更新日期2026-08-30
1 天星标数0
7 天星标数0
排名16,540
上次排名16,325
排名变化-215
7 天星标排名15,583
上次 7 天星标排名15,594
7 天星标排名变化11
总星标排名17,099
上次总星标排名17,080
总星标排名变化-19
README 摘录

Audio-Classification (Kapre Version) Pipeline for prototyping audio classification algorithms with TF 2. 3 YouTube Environment Jupyter Notebooks Audio Preprocessing Training Plot History Confusion Matrix Receiver Operating Characteristic Kapre YouTube This series has been re-worked. There are new videos to support this repository. It is recommended to follow the new series. If you want to follow the old videos, restore to a previous commit. Environment Jupyter Notebooks Assuming you have ipykernel installed from your conda environment Audio Preprocessing clean

. py can be used to preview the signal envelope at a threshold to remove low magnitude data When you uncomment split_wavs, a clean directory will be created with downsampled mono audio split by delta time Training Change model_type to: conv1d, conv2d, lstm Sample rate and delta time should be the same from clean. py Plot History Assuming you have ran all 3 models and saved the images into logs, check Confusion Matrix Receiver Operating Characteristic Kapre For computation

of audio transforms from time to frequency domain on the fly

常见问题

记录了多少星标和分叉？

seth814/Audio-Classification 有 586 个星标和 191 个分叉。

仓库排名是多少？

seth814/Audio-Classification 排名第 16,540。

记录的语言和类别是什么？

seth814/Audio-Classification 有 586 个星标，语言为 Jupyter Notebook，类别为 Tutorials。

哪个仓库排名最接近？

seth814/Audio-Classification 与 avidevelops/claude-architect-exam-prep 比较，后者排名第 16,533，有 586 个星标。

README

Audio-Classification (Kapre Version)

Pipeline for prototyping audio classification algorithms with TF 2.3

YouTube

This series has been re-worked. There are new videos to support this repository. It is recommended to follow the new series.

https://www.youtube.com/playlist?list=PLhA3b2k8R3t0SYW_MhWkWS5fWg-BlYqWn

If you want to follow the old videos, restore to a previous commit.

git checkout 404f2a6f989cec3421e8217d71ef070f3593a84d

Environment

conda create -n audio python=3.7
activate audio
pip install -r requirements.txt

Jupyter Notebooks

Assuming you have ipykernel installed from your conda environment

ipython kernel install --user --name=audio

conda activate audio

jupyter-notebook

Audio Preprocessing

clean.py can be used to preview the signal envelope at a threshold to remove low magnitude data

When you uncomment split_wavs, a clean directory will be created with downsampled mono audio split by delta time

python clean.py

Training

Change model_type to: conv1d, conv2d, lstm

Sample rate and delta time should be the same from clean.py

python train.py

Plot History

Assuming you have ran all 3 models and saved the images into logs, check notebooks/Plot History.ipynb

notebooks/Confusion Matrix and ROC.ipynb

Confusion Matrix

Receiver Operating Characteristic

Kapre

For computation of audio transforms from time to frequency domain on the fly

https://github.com/keunwoochoi/kapre
https://arxiv.org/pdf/1706.05781.pdf

排名相邻
同类项目