Upload folder using huggingface_hub
Browse files- TunSwitchCS.zip +3 -0
- TunSwitchTO.zip +3 -0
- language_annotation.zip +3 -0
- paper.pdf +0 -0
- readme.txt +19 -0
- test_wavs.zip +3 -0
- weakly_labeled_tn.zip +3 -0
TunSwitchCS.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22bb8aa0a7715e92a4187bd216cdd96d230c7716239f47843d8321af366dd8a4
|
| 3 |
+
size 1567483362
|
TunSwitchTO.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec53c56e1caf44fe0d13b5585c8fed5ad8e25677820372c8dcc9c8bf6d991e2b
|
| 3 |
+
size 92290570
|
language_annotation.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:858271576e30cd89aeb4ba3268d2f31c833d92f18f69f655802db97dfa3ab501
|
| 3 |
+
size 312136
|
paper.pdf
ADDED
|
Binary file (180 kB). View file
|
|
|
readme.txt
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
This folder contains the data used to develop and test the Tunisian Arabic Automatic Speech Recognition model developed in the following paper :
|
| 2 |
+
|
| 3 |
+
A. A. Ben Abdallah*, A. Kabboudi, A. Kanoun, and S. Zaiem*, “Leveraging data collection and
|
| 4 |
+
unsupervised learning for code-switched tunisian arabic automatic speech recognition”, Submitted to
|
| 5 |
+
ICASSP 2024, vol. * : These two authors have contributed equally. 2023.
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
It contains 4 zipped folders containing audio data :
|
| 9 |
+
- TunSwitchCS.zip : containing annotated code-switched data.
|
| 10 |
+
- TunSwitchTO.zip : containing annotated Tunisian-Only data.
|
| 11 |
+
- weakly_labeled_tn.zip : containing weakly-labeled (or unlabeled) audio data. Audios may contain code-switching, but the current weak labels do not.
|
| 12 |
+
- test_wavs.zip : contains annotated testing data, divided between a code-switched part and a tunisian-only part.
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
It also contains textual data, used for language modelling, contained in TextData.zip. Finally it also contains a language-detailed annotation of TunSwitchCS in the folder language_annotation/ .
|
| 16 |
+
|
| 17 |
+
More details about the data are available in the paper. The current table are in a SpeechBrain-friendly format, the column path is irrelevant and has to be changed according to your local setting. Please use the provided train-dev-test splits if you work with this dataset.
|
| 18 |
+
|
| 19 |
+
Please cite the aforementioned paper if you use or refer to this dataset.
|
test_wavs.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a8492897f10f40eba99797850fb539c627ac951dd39b40a26d40135c52c9096e
|
| 3 |
+
size 284095903
|
weakly_labeled_tn.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:adf3aa0241ff0b55651fa89e7297b3b9455adf6707f4f9e1a2c649874a8f910e
|
| 3 |
+
size 16041460441
|