基本用不了啊，好多依赖数据都没有，也没有生成逻辑。

[hanling6889580]

IEMOCAP_PATH = "data/iemocap.pickle" IEMOCAP_BALANCED_PATH = "data/iemocap_balanced.pickle" IEMOCAP_BALANCED_ASR_PATH = "data/iemocap_balanced_asr.pickle" IEMOCAP_FULL_PATH = "data/IEMOCAP_full_release" LINGUISTIC_DATASET_PATH = "data/linguistic_features.npy" LINGUISTIC_LABELS_PATH = "data/linguistic_labels.npy" ACOUSTIC_FEATURES_PATH = "data/acoustic_features.npy" ACOUSTIC_LABELS_PATH = "data/acoustic_labels.npy" LINGUISTIC_DATASET_ASR_PATH = "data/linguistic_features_asr.npy" LINGUISTIC_LABELS_ASR_PATH = "data/linguistic_labels_asr.npy" SPECTROGRAMS_FEATURES_PATH = "data/spectrograms_features.npy" SPECTROGRAMS_LABELS_PATH = "data/spectrograms_labels.npy" MAPPING_ID_TO_SAMPLE_PATH = "data/id_to_sample.json"

这些文件都是怎么生成的？？

[abdoulsn]

IEMOCAP_PATH = "data/iemocap.pickle" IEMOCAP_BALANCED_PATH = "data/iemocap_balanced.pickle" IEMOCAP_BALANCED_ASR_PATH = "data/iemocap_balanced_asr.pickle" IEMOCAP_FULL_PATH = "data/IEMOCAP_full_release" LINGUISTIC_DATASET_PATH = "data/linguistic_features.npy" LINGUISTIC_LABELS_PATH = "data/linguistic_labels.npy" ACOUSTIC_FEATURES_PATH = "data/acoustic_features.npy" ACOUSTIC_LABELS_PATH = "data/acoustic_labels.npy" LINGUISTIC_DATASET_ASR_PATH = "data/linguistic_features_asr.npy" LINGUISTIC_LABELS_ASR_PATH = "data/linguistic_labels_asr.npy" SPECTROGRAMS_FEATURES_PATH = "data/spectrograms_features.npy" SPECTROGRAMS_LABELS_PATH = "data/spectrograms_labels.npy" MAPPING_ID_TO_SAMPLE_PATH = "data/id_to_sample.json"

这些文件都是怎么生成的？？

What's this?

[hanling6889580]

这个文件是怎么生成的 ？iemocap_balanced_asr.pickle 代码里看不见生成这个文件的逻辑。

[PiotrSobczak]

Hi, I would appreciate english :) I will add info in docstrings to the module and let you know

[PiotrSobczak]

I added the comments to github.com/PiotrSobczak/speech-emotion-recognition/blob/master/speech_emotion_recognition/data_loader.py.

"""
1.Download IEMOCAP dataset from https://sail.usc.edu/iemocap/
2.Use github.com/didi/delta/blob/master/egs/iemocap/emo/v1/local/python/mocap_data_collect.py to get dataset pickle
3.Use create_balanced_iemocap() to get balanced version of iemocap dataset containing 4 classes
4.Use load_<DATASET_TYPE>_dataset to load a specific dataset. 
*The first time you use load functions it will be created from pickle. This might take a while...
*The next time you use load functions you will load cached .npy files for faster loading
"""

[PiotrSobczak]

I'm closing the issue, reopen if need more info. :)

[hanling6889580]

thanks ~

[PiotrSobczak]

@hanling6889580 I moved the info to README, turned out its quite important :D

[leijue222]

I added the comments to github.com/PiotrSobczak/speech-emotion-recognition/blob/master/speech_emotion_recognition/data_loader.py.

"""
1.Download IEMOCAP dataset from https://sail.usc.edu/iemocap/
2.Use github.com/didi/delta/blob/master/egs/iemocap/emo/v1/local/python/mocap_data_collect.py to get dataset pickle
3.Use create_balanced_iemocap() to get balanced version of iemocap dataset containing 4 classes
4.Use load_<DATASET_TYPE>_dataset to load a specific dataset. 
*The first time you use load functions it will be created from pickle. This might take a while...
*The next time you use load functions you will load cached .npy files for faster loading
"""

But a new problem is here: How to get the file of embeddings_array.numpy andword_to_index.pickle?

https://github.com/PiotrSobczak/speech-emotion-recognition/blob/8ecbcf11a6dbc6d6802e51f7f797fcd85fdb7992/speech_emotion_recognition/word2vec_wrapper.py#L14-L15

:sob: I have no any way to get them...