Convert incoming data list to tensor

[Korbrent]

Even though we don't know what features we want or how many, start taking a hypothetical output data from Preprocessing into tensors to prepare for the model

[Korbrent]

Guys I did some shit :smile:

[leosaa]

The WAVs dataset is available at:

https://www.cs.nmt.edu/~leo/CREMA-D/AudioWAV/

[leosaa]

be happy!

[leosaa]

the logic behind every file name:

"fileName" - name of the movie file rated
"VoiceVote" - the emotion (or emotions separated by a colon) with the majority vote for Voice ratings. (A, D, F, H, N, or S)
"VoiceLevel" - the numeric rating (or ratings separated by a colon) corresponding to the emotion(s) listed in "VoiceVote"
"FaceVote" - the emotion (or emotions separated by a colon) with the majority vote for Face ratings. (A, D, F, H, N, or S)
"FaceLevel" - the numeric rating (or ratings separated by a colon) corresponding to the emotion(s) listed in "FaceVote"
"MultiModalVote" - the emotion (or emotions separated by a colon) with the majority vote for MultiModal ratings. (A, D, F, H, N, or S)
"MultiModalLevel" - the numeric rating (or ratings separated by a colon) corresponding to the emotion(s) listed in "MultiModalVote"

[leosaa]

logic file names.

[leosaa]

The sentences were presented using different emotion (in parentheses is the three letter code used in the third part of the filename):

Anger (ANG)
Disgust (DIS)
Fear (FEA)
Happy/Joy (HAP)
Neutral (NEU)
Sad (SAD)

and emotion level (in parentheses is the two letter code used in the fourth part of the filename):

Low (LO)
Medium (MD)
High (HI)
Unspecified (XX)

[leosaa]

.

[Korbrent]

Reassigned to Leo. Fix the linker.py file to grab the .wav files from the online repository (hosted by @leosaa's cs page). The function prepare_dataset() should be modified for fetching the .wav files, and the function build_tensors() will need to be modified as well for the labels.

Also, the file has been changed from a .ipynb file into a .py file at @Catmaniscatlord's request.