Datasets:

Modalities:
Text
Formats:
csv
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
License:
Pendrokar commited on
Commit
5d352e4
Β·
verified Β·
1 Parent(s): b23fbb2

TTSDS Dataset legend

Browse files
Files changed (1) hide show
  1. README.md +30 -1
README.md CHANGED
@@ -106,6 +106,35 @@ Example if the proprietary ElevenLabs were to be added to the capabilities table
106
 
107
  More info on how the capabilities table came about can be found within the [GitHub Issue](https://github.com/Vaibhavs10/open-tts-tracker/issues/14).
108
 
109
- Please create pull requests to update the info on the models.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
110
 
111
  ---
 
106
 
107
  More info on how the capabilities table came about can be found within the [GitHub Issue](https://github.com/Vaibhavs10/open-tts-tracker/issues/14).
108
 
109
+ ## train_data Legend
110
+
111
+ Legend for the separate TTSDS Datasets ([_train_data_ viewer](https://huggingface.co/datasets/Pendrokar/open_tts_tracker/viewer/train_data) [GitHub](https://github.com/ttsds/ttsds_systems))
112
+
113
+ - 🌐 Multilingual
114
+ - The ISO codes of languages the model is capable off. ❌ if English only.
115
+ - πŸ“š Training Amount (k hours)
116
+ - The number of hours the model was trained on
117
+ - 🧠 Num. Parameters (M)
118
+ - How many parameters the model has, excluding vocoder and text-only components
119
+ - 🎯 Target Repr.
120
+ - Which output representations the model uses, for example audio codecs or mel spectrograms
121
+ - πŸ“– LibriVox Only
122
+ - If the model was trained on librivox-like (audiobook) data alone
123
+ - πŸ”„ NAR
124
+ - If the model has a significant non-autoregressive component
125
+ - πŸ” AR
126
+ - If the model has a significant autoregressive component
127
+ - πŸ”‘ G2P
128
+ - If the model uses G2P (phone inputs)
129
+ - 🧩 Language Model
130
+ - If an LM-like approach is used (next token prediction)
131
+ - 🎡 Prosody Prediction
132
+ - If prosodic correlates such as pitch or energy are predicted
133
+ - 🌊 Diffusion
134
+ - If diffusion is used (outside vocoder)
135
+ - ⏱️ Delay Pattern
136
+ - If a delay pattern is used for audio codes (see [Lyth & King, 2024](https://arxiv.org/abs/2402.01912))
137
+
138
+ _Please create pull requests to update the info on the models!_
139
 
140
  ---