kiranpantha
commited on
Commit
•
f97b638
1
Parent(s):
06a2158
End of training
Browse files
README.md
CHANGED
@@ -1,40 +1,42 @@
|
|
1 |
---
|
2 |
-
base_model: facebook/wav2vec2-xls-r-300m
|
3 |
-
datasets:
|
4 |
-
- arrow
|
5 |
library_name: transformers
|
|
|
|
|
6 |
license: apache-2.0
|
7 |
-
|
8 |
-
- wer
|
9 |
tags:
|
10 |
- generated_from_trainer
|
|
|
|
|
|
|
|
|
11 |
model-index:
|
12 |
-
- name:
|
13 |
results:
|
14 |
- task:
|
15 |
-
type: automatic-speech-recognition
|
16 |
name: Automatic Speech Recognition
|
|
|
17 |
dataset:
|
18 |
-
name:
|
19 |
-
type:
|
20 |
config: default
|
21 |
split: test
|
22 |
-
args:
|
23 |
metrics:
|
24 |
-
-
|
25 |
-
|
26 |
-
|
27 |
---
|
28 |
|
29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
30 |
should probably proofread and complete it, then remove this comment. -->
|
31 |
|
32 |
-
#
|
33 |
|
34 |
-
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the
|
35 |
It achieves the following results on the evaluation set:
|
36 |
-
- Loss: 0.
|
37 |
-
- Wer: 0.
|
38 |
|
39 |
## Model description
|
40 |
|
@@ -53,66 +55,52 @@ More information needed
|
|
53 |
### Training hyperparameters
|
54 |
|
55 |
The following hyperparameters were used during training:
|
56 |
-
- learning_rate:
|
57 |
- train_batch_size: 8
|
58 |
- eval_batch_size: 8
|
59 |
- seed: 42
|
60 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
61 |
- lr_scheduler_type: linear
|
62 |
- lr_scheduler_warmup_steps: 500
|
63 |
-
- num_epochs:
|
64 |
- mixed_precision_training: Native AMP
|
65 |
|
66 |
### Training results
|
67 |
|
68 |
-
| Training Loss | Epoch | Step
|
69 |
-
|
70 |
-
|
|
71 |
-
| 2.
|
72 |
-
| 1.
|
73 |
-
| 1.
|
74 |
-
| 0.
|
75 |
-
| 0.
|
76 |
-
| 0.
|
77 |
-
| 0.
|
78 |
-
| 0.
|
79 |
-
| 0.
|
80 |
-
| 0.
|
81 |
-
| 0.
|
82 |
-
| 0.
|
83 |
-
| 0.
|
84 |
-
| 0.
|
85 |
-
| 0.
|
86 |
-
| 0.
|
87 |
-
| 0.
|
88 |
-
| 0.
|
89 |
-
| 0.
|
90 |
-
| 0.
|
91 |
-
| 0.
|
92 |
-
| 0.
|
93 |
-
| 0.
|
94 |
-
| 0.
|
95 |
-
| 0.
|
96 |
-
| 0.
|
97 |
-
| 0.2621 | 6.7187 | 11200 | 0.4119 | 0.4211 |
|
98 |
-
| 0.2632 | 6.9586 | 11600 | 0.4037 | 0.4266 |
|
99 |
-
| 0.2201 | 7.1986 | 12000 | 0.4239 | 0.4229 |
|
100 |
-
| 0.2215 | 7.4385 | 12400 | 0.4266 | 0.4213 |
|
101 |
-
| 0.213 | 7.6785 | 12800 | 0.4149 | 0.4229 |
|
102 |
-
| 0.202 | 7.9184 | 13200 | 0.4154 | 0.4113 |
|
103 |
-
| 0.2034 | 8.1584 | 13600 | 0.4193 | 0.4081 |
|
104 |
-
| 0.1809 | 8.3983 | 14000 | 0.4164 | 0.4090 |
|
105 |
-
| 0.185 | 8.6383 | 14400 | 0.4154 | 0.3942 |
|
106 |
-
| 0.1813 | 8.8782 | 14800 | 0.4078 | 0.3914 |
|
107 |
-
| 0.1625 | 9.1182 | 15200 | 0.4223 | 0.3956 |
|
108 |
-
| 0.1642 | 9.3581 | 15600 | 0.4204 | 0.3926 |
|
109 |
-
| 0.1561 | 9.5981 | 16000 | 0.4172 | 0.3880 |
|
110 |
-
| 0.155 | 9.8380 | 16400 | 0.4180 | 0.3861 |
|
111 |
|
112 |
|
113 |
### Framework versions
|
114 |
|
115 |
- Transformers 4.45.0.dev0
|
116 |
-
- Pytorch 2.4.
|
117 |
- Datasets 2.21.0
|
118 |
- Tokenizers 0.19.1
|
|
|
1 |
---
|
|
|
|
|
|
|
2 |
library_name: transformers
|
3 |
+
language:
|
4 |
+
- ne
|
5 |
license: apache-2.0
|
6 |
+
base_model: facebook/wav2vec2-xls-r-300m
|
|
|
7 |
tags:
|
8 |
- generated_from_trainer
|
9 |
+
datasets:
|
10 |
+
- kiranpantha/OpenSLR54-Balanced-Nepali
|
11 |
+
metrics:
|
12 |
+
- wer
|
13 |
model-index:
|
14 |
+
- name: XLSR-300M-Nepali
|
15 |
results:
|
16 |
- task:
|
|
|
17 |
name: Automatic Speech Recognition
|
18 |
+
type: automatic-speech-recognition
|
19 |
dataset:
|
20 |
+
name: OpenSLR54
|
21 |
+
type: kiranpantha/OpenSLR54-Balanced-Nepali
|
22 |
config: default
|
23 |
split: test
|
24 |
+
args: 'config: ne, split: train,test'
|
25 |
metrics:
|
26 |
+
- name: Wer
|
27 |
+
type: wer
|
28 |
+
value: 0.5244204160175937
|
29 |
---
|
30 |
|
31 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
32 |
should probably proofread and complete it, then remove this comment. -->
|
33 |
|
34 |
+
# XLSR-300M-Nepali
|
35 |
|
36 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the OpenSLR54 dataset.
|
37 |
It achieves the following results on the evaluation set:
|
38 |
+
- Loss: 0.2681
|
39 |
+
- Wer: 0.5244
|
40 |
|
41 |
## Model description
|
42 |
|
|
|
55 |
### Training hyperparameters
|
56 |
|
57 |
The following hyperparameters were used during training:
|
58 |
+
- learning_rate: 5e-05
|
59 |
- train_batch_size: 8
|
60 |
- eval_batch_size: 8
|
61 |
- seed: 42
|
62 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
63 |
- lr_scheduler_type: linear
|
64 |
- lr_scheduler_warmup_steps: 500
|
65 |
+
- num_epochs: 2
|
66 |
- mixed_precision_training: Native AMP
|
67 |
|
68 |
### Training results
|
69 |
|
70 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
71 |
+
|:-------------:|:------:|:----:|:---------------:|:------:|
|
72 |
+
| 3.2642 | 0.0722 | 300 | 2.9627 | 1.0 |
|
73 |
+
| 2.1949 | 0.1444 | 600 | 1.5526 | 1.0160 |
|
74 |
+
| 1.4595 | 0.2166 | 900 | 1.1674 | 0.9810 |
|
75 |
+
| 1.2128 | 0.2888 | 1200 | 0.9901 | 0.9668 |
|
76 |
+
| 0.976 | 0.3610 | 1500 | 0.6942 | 0.7696 |
|
77 |
+
| 0.8267 | 0.4332 | 1800 | 0.6314 | 0.7552 |
|
78 |
+
| 0.7542 | 0.5054 | 2100 | 0.5522 | 0.7156 |
|
79 |
+
| 0.7228 | 0.5776 | 2400 | 0.5210 | 0.6960 |
|
80 |
+
| 0.6707 | 0.6498 | 2700 | 0.4744 | 0.6581 |
|
81 |
+
| 0.6368 | 0.7220 | 3000 | 0.4529 | 0.6535 |
|
82 |
+
| 0.5944 | 0.7942 | 3300 | 0.4229 | 0.6264 |
|
83 |
+
| 0.5651 | 0.8664 | 3600 | 0.4061 | 0.6161 |
|
84 |
+
| 0.5469 | 0.9386 | 3900 | 0.3788 | 0.6103 |
|
85 |
+
| 0.5308 | 1.0108 | 4200 | 0.3668 | 0.5957 |
|
86 |
+
| 0.4684 | 1.0830 | 4500 | 0.3509 | 0.5920 |
|
87 |
+
| 0.4382 | 1.1552 | 4800 | 0.3398 | 0.5920 |
|
88 |
+
| 0.4424 | 1.2274 | 5100 | 0.3260 | 0.5767 |
|
89 |
+
| 0.4159 | 1.2996 | 5400 | 0.3189 | 0.5690 |
|
90 |
+
| 0.419 | 1.3718 | 5700 | 0.3067 | 0.5581 |
|
91 |
+
| 0.4114 | 1.4440 | 6000 | 0.3019 | 0.5568 |
|
92 |
+
| 0.3903 | 1.5162 | 6300 | 0.2982 | 0.5549 |
|
93 |
+
| 0.3915 | 1.5884 | 6600 | 0.2887 | 0.5493 |
|
94 |
+
| 0.3789 | 1.6606 | 6900 | 0.2813 | 0.5398 |
|
95 |
+
| 0.3725 | 1.7329 | 7200 | 0.2763 | 0.5339 |
|
96 |
+
| 0.3706 | 1.8051 | 7500 | 0.2704 | 0.5285 |
|
97 |
+
| 0.3624 | 1.8773 | 7800 | 0.2706 | 0.5264 |
|
98 |
+
| 0.357 | 1.9495 | 8100 | 0.2681 | 0.5244 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
99 |
|
100 |
|
101 |
### Framework versions
|
102 |
|
103 |
- Transformers 4.45.0.dev0
|
104 |
+
- Pytorch 2.4.1+cu121
|
105 |
- Datasets 2.21.0
|
106 |
- Tokenizers 0.19.1
|