EmoTag helps you train emotion detection model for Chinese audios

Last update: Sep 07, 2022

Overview

emoTag

emoTag helps you train emotion detection model for Chinese audios.

Environment

pip install -r requirement.txt

Data

We used Emotional Speech Dataset (ESD) for Speech Synthesis and Voice Conversion from HLT Singapore.

Train Emotion Classifier

Use this command to train a classifier. Adjust training setups in conf/logfbank_train-emo.json.

python train.py --config conf/logfbank_train-emo.json --name task_trial_1

Models and logs will be find in exp/.

usage: train.py [-h] [-c CONFIG] [-r RESUME] [-n NAME] [--lr LR] [--bs BS]
                [--train_utt2wav TRAIN_UTT2WAV] [--val_utt2wav VAL_UTT2WAV]
                [--blocks BLOCKS] [--optimizer OPTIMIZER]
                [--train_pad0 TRAIN_PAD0] [--devel_pad0 DEVEL_PAD0]
                [--pretrain PRETRAIN]

PyTorch Template

optional arguments:
  -h, --help            show this help message and exit
  -c CONFIG, --config CONFIG
                        config file path (default: None)
  -r RESUME, --resume RESUME
                        path to latest checkpoint (default: None)
  -n NAME, --name NAME
  --lr LR, --learning_rate LR
  --bs BS, --batch_size BS
  --train_utt2wav TRAIN_UTT2WAV
  --val_utt2wav VAL_UTT2WAV
  --blocks BLOCKS
  --optimizer OPTIMIZER
  --train_pad0 TRAIN_PAD0
  --devel_pad0 DEVEL_PAD0
  --pretrain PRETRAIN

Infer labels

python infer_label.py

Adjust the vad_file param and code if necessary to adapt to new tasks. infer_label.py adopted multiprocessing, increased cpu utilities rate and inference efficiency. See usage details below.

usage: infer_label.py [-h] [--vad_file VAD_FILE] [--model_dir MODEL_DIR]
                      [--output_dir OUTPUT_DIR]

parse model info

optional arguments:
  -h, --help            show this help message and exit
  --vad_file VAD_FILE
  --model_dir MODEL_DIR
  --output_dir OUTPUT_DIR

EmoTag helps you train emotion detection model for Chinese audios

Related tags

Overview

emoTag

Environment

Data

Train Emotion Classifier

Infer labels

Owner

_zza

HyperLib: Deep learning in the Hyperbolic space

unofficial pytorch implement of "Squareplus: A Softplus-Like Algebraic Rectifier"

Exploring Versatile Prior for Human Motion via Motion Frequency Guidance (3DV2021)

Official Pytorch Implementation of Length-Adaptive Transformer (ACL 2021)

Hand tracking demo for DIY Smart Glasses with a remote computer doing the work

Library to enable Bayesian active learning in your research or labeling work.

Official repo for QHack—the quantum machine learning hackathon

Code for Transformers Solve Limited Receptive Field for Monocular Depth Prediction

A denoising diffusion probabilistic model synthesises galaxies that are qualitatively and physically indistinguishable from the real thing.

Generalizing Gaze Estimation with Outlier-guided Collaborative Adaptation

CPU inference engine that delivers unprecedented performance for sparse models

Transformer model implemented with Pytorch

Prompt Tuning with Rules

Code for the SIGGRAPH 2021 paper "Consistent Depth of Moving Objects in Video".

Using python and scikit-learn to make stock predictions

An easier way to build neural search on the cloud

Tree Nested PyTorch Tensor Lib

Code for 2021 NeurIPS --- Towards Multi-Grained Explainability for Graph Neural Networks

Setup freqtrade/freqUI on Heroku

PyTorch implementation for the visual prior component (i.e. perception module) of the Visually Grounded Physics Learner [Li et al., 2020].