Update README.md
Browse files
README.md
CHANGED
|
@@ -23,9 +23,52 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 23 |
|
| 24 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the GOOGLE/XTREME_S - MINDS14.ALL dataset.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
-
-
|
| 27 |
-
-
|
| 28 |
-
- Accuracy: 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
|
|
| 23 |
|
| 24 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the GOOGLE/XTREME_S - MINDS14.ALL dataset.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
+
- Accuracy: 0.9033
|
| 27 |
+
- Accuracy Cs-cz: 0.9164
|
| 28 |
+
- Accuracy De-de: 0.9477
|
| 29 |
+
- Accuracy En-au: 0.9235
|
| 30 |
+
- Accuracy En-gb: 0.9324
|
| 31 |
+
- Accuracy En-us: 0.9326
|
| 32 |
+
- Accuracy Es-es: 0.9177
|
| 33 |
+
- Accuracy Fr-fr: 0.9444
|
| 34 |
+
- Accuracy It-it: 0.9167
|
| 35 |
+
- Accuracy Ko-kr: 0.8649
|
| 36 |
+
- Accuracy Nl-nl: 0.9450
|
| 37 |
+
- Accuracy Pl-pl: 0.9146
|
| 38 |
+
- Accuracy Pt-pt: 0.8940
|
| 39 |
+
- Accuracy Ru-ru: 0.8667
|
| 40 |
+
- Accuracy Zh-cn: 0.7291
|
| 41 |
+
- F1: 0.9015
|
| 42 |
+
- F1 Cs-cz: 0.9154
|
| 43 |
+
- F1 De-de: 0.9467
|
| 44 |
+
- F1 En-au: 0.9199
|
| 45 |
+
- F1 En-gb: 0.9334
|
| 46 |
+
- F1 En-us: 0.9308
|
| 47 |
+
- F1 Es-es: 0.9158
|
| 48 |
+
- F1 Fr-fr: 0.9436
|
| 49 |
+
- F1 It-it: 0.9135
|
| 50 |
+
- F1 Ko-kr: 0.8642
|
| 51 |
+
- F1 Nl-nl: 0.9440
|
| 52 |
+
- F1 Pl-pl: 0.9159
|
| 53 |
+
- F1 Pt-pt: 0.8883
|
| 54 |
+
- F1 Ru-ru: 0.8646
|
| 55 |
+
- F1 Zh-cn: 0.7249
|
| 56 |
+
- Loss: 0.4119
|
| 57 |
+
- Loss Cs-cz: 0.3790
|
| 58 |
+
- Loss De-de: 0.2649
|
| 59 |
+
- Loss En-au: 0.3459
|
| 60 |
+
- Loss En-gb: 0.2853
|
| 61 |
+
- Loss En-us: 0.2203
|
| 62 |
+
- Loss Es-es: 0.2731
|
| 63 |
+
- Loss Fr-fr: 0.1909
|
| 64 |
+
- Loss It-it: 0.3520
|
| 65 |
+
- Loss Ko-kr: 0.5431
|
| 66 |
+
- Loss Nl-nl: 0.2515
|
| 67 |
+
- Loss Pl-pl: 0.4113
|
| 68 |
+
- Loss Pt-pt: 0.4798
|
| 69 |
+
- Loss Ru-ru: 0.6470
|
| 70 |
+
- Loss Zh-cn: 1.1216
|
| 71 |
+
- Predict Samples: 4086
|
| 72 |
|
| 73 |
## Model description
|
| 74 |
|