bagustris / dimensional-ser Goto Github PK

Repository for my paper: Dimensional Speech Emotion Recognition Using Acoustic Features and Word Embeddings using Multitask Learning

Home Page: https://www.cambridge.org/core/journals/apsipa-transactions-on-signal-and-information-processing/article/dimensional-speech-emotion-recognition-from-speech-features-and-word-embeddings-by-using-multitask-learning/BCBF69FBED76857F84090A2FB58B2498

Python 100.00%

speech-emotion-recognition feature-fusion dimensional-emotion

dimensional-ser's Issues

Word embeddings

Great paper. Do you have code for word embeddings acquisition (i.e. that was used to generate x_train_text.npy and g_word_embedding_matrix.npy)? I'm also working with the IEMOCAP dataset and this would be very helpful. Thanks! @bagustris

About text features

Could you plz provide processing text transcription py.file? :)
Such as the file to produce x_train_text.npy
g_word_embedding_matrix.npy
crawl-300d-2M-subword.vec
I use word2vec to transfrom text to vector and not sure if it's correct.

About dimension of features and vad labels

plz use hst.npy for features and keep the dimension of features as (10039,*)and vad as (10039,3),you can cancel the code :vad= vad.reshape(-1,1)

Request for SEWA Dataset

Could you plz provide the SEWA database? We saw your paper in 2020ICASSP used the dataset. We have sent the request to SEWA team, but didn't get reply. We only use it for academic research.

Recommend Projects