AndrewMcDowell
commited on
Commit
•
d448748
1
Parent(s):
0c58e25
update model card README.md
Browse files
README.md
CHANGED
@@ -1,10 +1,6 @@
|
|
1 |
---
|
2 |
-
language:
|
3 |
-
- ar
|
4 |
license: apache-2.0
|
5 |
tags:
|
6 |
-
- automatic-speech-recognition
|
7 |
-
- mozilla-foundation/common_voice_8_0
|
8 |
- generated_from_trainer
|
9 |
datasets:
|
10 |
- common_voice
|
@@ -18,10 +14,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
18 |
|
19 |
#
|
20 |
|
21 |
-
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the
|
22 |
It achieves the following results on the evaluation set:
|
23 |
-
- Loss: 1.
|
24 |
-
- Wer: 0.
|
25 |
|
26 |
## Model description
|
27 |
|
@@ -49,24 +45,48 @@ The following hyperparameters were used during training:
|
|
49 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
50 |
- lr_scheduler_type: linear
|
51 |
- lr_scheduler_warmup_steps: 2000
|
52 |
-
- num_epochs:
|
53 |
- mixed_precision_training: Native AMP
|
54 |
|
55 |
### Training results
|
56 |
|
57 |
-
| Training Loss | Epoch | Step
|
58 |
-
|
59 |
-
| 2.2416 | 0.84 | 500
|
60 |
-
| 2.3089 | 1.67 | 1000
|
61 |
-
| 2.3614 | 2.51 | 1500
|
62 |
-
| 2.5234 | 3.35 | 2000
|
63 |
-
| 2.5373 | 4.19 | 2500
|
64 |
-
| 2.5703 | 5.03 | 3000
|
65 |
-
| 2.4656 | 5.86 | 3500
|
66 |
-
| 2.4339 | 6.7 | 4000
|
67 |
-
| 2.344 | 7.54 | 4500
|
68 |
-
| 2.2677 | 8.38 | 5000
|
69 |
-
| 2.2074 | 9.21 | 5500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
70 |
|
71 |
|
72 |
### Framework versions
|
|
|
1 |
---
|
|
|
|
|
2 |
license: apache-2.0
|
3 |
tags:
|
|
|
|
|
4 |
- generated_from_trainer
|
5 |
datasets:
|
6 |
- common_voice
|
|
|
14 |
|
15 |
#
|
16 |
|
17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the common_voice dataset.
|
18 |
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 1.1414
|
20 |
+
- Wer: 0.8616
|
21 |
|
22 |
## Model description
|
23 |
|
|
|
45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
- lr_scheduler_type: linear
|
47 |
- lr_scheduler_warmup_steps: 2000
|
48 |
+
- num_epochs: 30.0
|
49 |
- mixed_precision_training: Native AMP
|
50 |
|
51 |
### Training results
|
52 |
|
53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
54 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
55 |
+
| 2.2416 | 0.84 | 500 | 1.2867 | 0.8875 |
|
56 |
+
| 2.3089 | 1.67 | 1000 | 1.8336 | 0.9548 |
|
57 |
+
| 2.3614 | 2.51 | 1500 | 1.5937 | 0.9469 |
|
58 |
+
| 2.5234 | 3.35 | 2000 | 1.9765 | 0.9867 |
|
59 |
+
| 2.5373 | 4.19 | 2500 | 1.9062 | 0.9916 |
|
60 |
+
| 2.5703 | 5.03 | 3000 | 1.9772 | 0.9915 |
|
61 |
+
| 2.4656 | 5.86 | 3500 | 1.8083 | 0.9829 |
|
62 |
+
| 2.4339 | 6.7 | 4000 | 1.7548 | 0.9752 |
|
63 |
+
| 2.344 | 7.54 | 4500 | 1.6146 | 0.9638 |
|
64 |
+
| 2.2677 | 8.38 | 5000 | 1.5105 | 0.9499 |
|
65 |
+
| 2.2074 | 9.21 | 5500 | 1.4191 | 0.9357 |
|
66 |
+
| 2.3768 | 10.05 | 6000 | 1.6663 | 0.9665 |
|
67 |
+
| 2.3804 | 10.89 | 6500 | 1.6571 | 0.9720 |
|
68 |
+
| 2.3237 | 11.72 | 7000 | 1.6049 | 0.9637 |
|
69 |
+
| 2.317 | 12.56 | 7500 | 1.5875 | 0.9655 |
|
70 |
+
| 2.2988 | 13.4 | 8000 | 1.5357 | 0.9603 |
|
71 |
+
| 2.2906 | 14.24 | 8500 | 1.5637 | 0.9592 |
|
72 |
+
| 2.2848 | 15.08 | 9000 | 1.5326 | 0.9537 |
|
73 |
+
| 2.2381 | 15.91 | 9500 | 1.5631 | 0.9508 |
|
74 |
+
| 2.2072 | 16.75 | 10000 | 1.4565 | 0.9395 |
|
75 |
+
| 2.197 | 17.59 | 10500 | 1.4304 | 0.9406 |
|
76 |
+
| 2.198 | 18.43 | 11000 | 1.4230 | 0.9382 |
|
77 |
+
| 2.1668 | 19.26 | 11500 | 1.3998 | 0.9315 |
|
78 |
+
| 2.1498 | 20.1 | 12000 | 1.3920 | 0.9258 |
|
79 |
+
| 2.1244 | 20.94 | 12500 | 1.3584 | 0.9153 |
|
80 |
+
| 2.0953 | 21.78 | 13000 | 1.3274 | 0.9054 |
|
81 |
+
| 2.0762 | 22.61 | 13500 | 1.2933 | 0.9073 |
|
82 |
+
| 2.0587 | 23.45 | 14000 | 1.2516 | 0.8944 |
|
83 |
+
| 2.0363 | 24.29 | 14500 | 1.2214 | 0.8902 |
|
84 |
+
| 2.0302 | 25.13 | 15000 | 1.2087 | 0.8871 |
|
85 |
+
| 2.0071 | 25.96 | 15500 | 1.1953 | 0.8786 |
|
86 |
+
| 1.9882 | 26.8 | 16000 | 1.1738 | 0.8712 |
|
87 |
+
| 1.9772 | 27.64 | 16500 | 1.1647 | 0.8672 |
|
88 |
+
| 1.9585 | 28.48 | 17000 | 1.1459 | 0.8635 |
|
89 |
+
| 1.944 | 29.31 | 17500 | 1.1414 | 0.8616 |
|
90 |
|
91 |
|
92 |
### Framework versions
|