update model card README.md
Browse files
README.md
CHANGED
@@ -2,11 +2,26 @@
|
|
2 |
license: apache-2.0
|
3 |
tags:
|
4 |
- generated_from_trainer
|
|
|
|
|
5 |
metrics:
|
6 |
- wer
|
7 |
model-index:
|
8 |
- name: wav2vec2LugandaASR
|
9 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
---
|
11 |
|
12 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -14,10 +29,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
14 |
|
15 |
# wav2vec2LugandaASR
|
16 |
|
17 |
-
This model is a fine-tuned version of [
|
18 |
It achieves the following results on the evaluation set:
|
19 |
-
- Loss: 0.
|
20 |
-
- Wer: 0.
|
21 |
|
22 |
## Model description
|
23 |
|
@@ -37,86 +52,47 @@ More information needed
|
|
37 |
|
38 |
The following hyperparameters were used during training:
|
39 |
- learning_rate: 0.0003
|
40 |
-
- train_batch_size:
|
41 |
- eval_batch_size: 8
|
42 |
- seed: 42
|
43 |
-
- gradient_accumulation_steps:
|
44 |
-
- total_train_batch_size:
|
45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
- lr_scheduler_type: linear
|
47 |
-
- lr_scheduler_warmup_steps:
|
48 |
-
- num_epochs:
|
49 |
|
50 |
### Training results
|
51 |
|
52 |
-
| Training Loss | Epoch
|
53 |
-
|
54 |
-
|
|
55 |
-
|
|
56 |
-
|
|
57 |
-
| 0.
|
58 |
-
| 0.
|
59 |
-
| 0.
|
60 |
-
| 0.
|
61 |
-
| 0.
|
62 |
-
| 0.
|
63 |
-
| 0.
|
64 |
-
| 0.
|
65 |
-
| 0.
|
66 |
-
| 0.
|
67 |
-
| 0.
|
68 |
-
| 0.
|
69 |
-
| 0.
|
70 |
-
| 0.
|
71 |
-
| 0.
|
72 |
-
| 0.
|
73 |
-
| 0.
|
74 |
-
| 0.
|
75 |
-
| 0.
|
76 |
-
| 0.
|
77 |
-
| 0.
|
78 |
-
| 0.
|
79 |
-
| 0.
|
80 |
-
| 0.
|
81 |
-
| 0.0554 | 54.37 | 2800 | 0.6340 | 0.5555 |
|
82 |
-
| 0.0515 | 56.31 | 2900 | 0.6500 | 0.5762 |
|
83 |
-
| 0.0439 | 58.25 | 3000 | 0.6376 | 0.5758 |
|
84 |
-
| 0.0461 | 60.19 | 3100 | 0.6265 | 0.5711 |
|
85 |
-
| 0.0479 | 62.14 | 3200 | 0.6230 | 0.5707 |
|
86 |
-
| 0.039 | 64.08 | 3300 | 0.6337 | 0.5584 |
|
87 |
-
| 0.0397 | 66.02 | 3400 | 0.6347 | 0.5736 |
|
88 |
-
| 0.0509 | 67.96 | 3500 | 0.5946 | 0.5483 |
|
89 |
-
| 0.0471 | 69.9 | 3600 | 0.6355 | 0.5584 |
|
90 |
-
| 0.0481 | 71.84 | 3700 | 0.6514 | 0.5559 |
|
91 |
-
| 0.0484 | 73.79 | 3800 | 0.6373 | 0.5566 |
|
92 |
-
| 0.041 | 75.73 | 3900 | 0.6736 | 0.5646 |
|
93 |
-
| 0.0349 | 77.67 | 4000 | 0.6375 | 0.5622 |
|
94 |
-
| 0.0349 | 79.61 | 4100 | 0.6158 | 0.5506 |
|
95 |
-
| 0.0273 | 81.55 | 4200 | 0.6914 | 0.5666 |
|
96 |
-
| 0.029 | 83.5 | 4300 | 0.6361 | 0.5399 |
|
97 |
-
| 0.0353 | 85.44 | 4400 | 0.6397 | 0.5584 |
|
98 |
-
| 0.0289 | 87.38 | 4500 | 0.6554 | 0.5499 |
|
99 |
-
| 0.0257 | 89.32 | 4600 | 0.6676 | 0.5557 |
|
100 |
-
| 0.0403 | 91.26 | 4700 | 0.6440 | 0.5584 |
|
101 |
-
| 0.0361 | 93.2 | 4800 | 0.6587 | 0.5521 |
|
102 |
-
| 0.0304 | 95.15 | 4900 | 0.6837 | 0.5454 |
|
103 |
-
| 0.0289 | 97.09 | 5000 | 0.6684 | 0.5370 |
|
104 |
-
| 0.0282 | 99.03 | 5100 | 0.6556 | 0.5296 |
|
105 |
-
| 0.0302 | 100.97 | 5200 | 0.6833 | 0.5394 |
|
106 |
-
| 0.0196 | 102.91 | 5300 | 0.6837 | 0.5291 |
|
107 |
-
| 0.0255 | 104.85 | 5400 | 0.6644 | 0.5374 |
|
108 |
-
| 0.0209 | 106.8 | 5500 | 0.6700 | 0.5289 |
|
109 |
-
| 0.0243 | 108.74 | 5600 | 0.6835 | 0.5338 |
|
110 |
-
| 0.0203 | 110.68 | 5700 | 0.6850 | 0.5410 |
|
111 |
-
| 0.0237 | 112.62 | 5800 | 0.6561 | 0.5349 |
|
112 |
-
| 0.0251 | 114.56 | 5900 | 0.6776 | 0.5298 |
|
113 |
-
| 0.0177 | 116.5 | 6000 | 0.6748 | 0.5282 |
|
114 |
-
| 0.0232 | 118.45 | 6100 | 0.6767 | 0.5296 |
|
115 |
-
| 0.0257 | 120.39 | 6200 | 0.6793 | 0.5320 |
|
116 |
-
| 0.0194 | 122.33 | 6300 | 0.6804 | 0.5303 |
|
117 |
-
| 0.0304 | 124.27 | 6400 | 0.6798 | 0.5287 |
|
118 |
-
| 0.0251 | 126.21 | 6500 | 0.6798 | 0.5291 |
|
119 |
-
| 0.0201 | 128.16 | 6600 | 0.6798 | 0.5291 |
|
120 |
|
121 |
|
122 |
### Framework versions
|
|
|
2 |
license: apache-2.0
|
3 |
tags:
|
4 |
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- common_voice_13_0
|
7 |
metrics:
|
8 |
- wer
|
9 |
model-index:
|
10 |
- name: wav2vec2LugandaASR
|
11 |
+
results:
|
12 |
+
- task:
|
13 |
+
name: Automatic Speech Recognition
|
14 |
+
type: automatic-speech-recognition
|
15 |
+
dataset:
|
16 |
+
name: common_voice_13_0
|
17 |
+
type: common_voice_13_0
|
18 |
+
config: lg
|
19 |
+
split: validation
|
20 |
+
args: lg
|
21 |
+
metrics:
|
22 |
+
- name: Wer
|
23 |
+
type: wer
|
24 |
+
value: 0.23959817157435953
|
25 |
---
|
26 |
|
27 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
29 |
|
30 |
# wav2vec2LugandaASR
|
31 |
|
32 |
+
This model is a fine-tuned version of [Gemmar/wav2vec2LugandaASR](https://huggingface.co/Gemmar/wav2vec2LugandaASR) on the common_voice_13_0 dataset.
|
33 |
It achieves the following results on the evaluation set:
|
34 |
+
- Loss: 0.2014
|
35 |
+
- Wer: 0.2396
|
36 |
|
37 |
## Model description
|
38 |
|
|
|
52 |
|
53 |
The following hyperparameters were used during training:
|
54 |
- learning_rate: 0.0003
|
55 |
+
- train_batch_size: 32
|
56 |
- eval_batch_size: 8
|
57 |
- seed: 42
|
58 |
+
- gradient_accumulation_steps: 4
|
59 |
+
- total_train_batch_size: 128
|
60 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
61 |
- lr_scheduler_type: linear
|
62 |
+
- lr_scheduler_warmup_steps: 200
|
63 |
+
- num_epochs: 5
|
64 |
|
65 |
### Training results
|
66 |
|
67 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
68 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
69 |
+
| 5.8963 | 0.18 | 100 | 2.8825 | 1.0000 |
|
70 |
+
| 1.1814 | 0.36 | 200 | 0.3787 | 0.4585 |
|
71 |
+
| 0.3331 | 0.54 | 300 | 0.3166 | 0.3918 |
|
72 |
+
| 0.2939 | 0.72 | 400 | 0.2811 | 0.3483 |
|
73 |
+
| 0.2682 | 0.9 | 500 | 0.2652 | 0.3348 |
|
74 |
+
| 0.2389 | 1.08 | 600 | 0.2565 | 0.3207 |
|
75 |
+
| 0.2137 | 1.27 | 700 | 0.2452 | 0.3066 |
|
76 |
+
| 0.2062 | 1.45 | 800 | 0.2356 | 0.3092 |
|
77 |
+
| 0.2058 | 1.63 | 900 | 0.2346 | 0.2928 |
|
78 |
+
| 0.2055 | 1.81 | 1000 | 0.2252 | 0.2901 |
|
79 |
+
| 0.1979 | 1.99 | 1100 | 0.2215 | 0.2836 |
|
80 |
+
| 0.166 | 2.17 | 1200 | 0.2217 | 0.2811 |
|
81 |
+
| 0.1623 | 2.35 | 1300 | 0.2200 | 0.2685 |
|
82 |
+
| 0.1628 | 2.53 | 1400 | 0.2166 | 0.2707 |
|
83 |
+
| 0.1593 | 2.71 | 1500 | 0.2131 | 0.2634 |
|
84 |
+
| 0.1561 | 2.89 | 1600 | 0.2121 | 0.2661 |
|
85 |
+
| 0.146 | 3.07 | 1700 | 0.2128 | 0.2552 |
|
86 |
+
| 0.1339 | 3.25 | 1800 | 0.2119 | 0.2591 |
|
87 |
+
| 0.1314 | 3.43 | 1900 | 0.2090 | 0.2492 |
|
88 |
+
| 0.1296 | 3.62 | 2000 | 0.2058 | 0.2504 |
|
89 |
+
| 0.1304 | 3.8 | 2100 | 0.2057 | 0.2500 |
|
90 |
+
| 0.1276 | 3.98 | 2200 | 0.2028 | 0.2463 |
|
91 |
+
| 0.116 | 4.16 | 2300 | 0.2058 | 0.2461 |
|
92 |
+
| 0.1122 | 4.34 | 2400 | 0.2074 | 0.2443 |
|
93 |
+
| 0.1087 | 4.52 | 2500 | 0.2065 | 0.2411 |
|
94 |
+
| 0.1087 | 4.7 | 2600 | 0.2042 | 0.2412 |
|
95 |
+
| 0.11 | 4.88 | 2700 | 0.2014 | 0.2396 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
96 |
|
97 |
|
98 |
### Framework versions
|