Model save
Browse files
README.md
CHANGED
@@ -1,25 +1,40 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
|
|
3 |
tags:
|
4 |
- generated_from_trainer
|
5 |
-
|
|
|
6 |
metrics:
|
7 |
- wer
|
8 |
model-index:
|
9 |
-
- name: wav2vec2-large-xlsr-53-
|
10 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
---
|
12 |
|
13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
14 |
should probably proofread and complete it, then remove this comment. -->
|
15 |
|
16 |
-
# wav2vec2-large-xlsr-53-
|
17 |
|
18 |
-
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 0.
|
21 |
-
- Wer:
|
22 |
-
- Cer:
|
23 |
|
24 |
## Model description
|
25 |
|
@@ -39,45 +54,39 @@ More information needed
|
|
39 |
|
40 |
The following hyperparameters were used during training:
|
41 |
- learning_rate: 6e-05
|
42 |
-
- train_batch_size:
|
43 |
- eval_batch_size: 8
|
44 |
- seed: 42
|
45 |
-
- gradient_accumulation_steps:
|
46 |
- total_train_batch_size: 32
|
47 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
48 |
- lr_scheduler_type: linear
|
49 |
-
-
|
50 |
-
- num_epochs:
|
51 |
- mixed_precision_training: Native AMP
|
52 |
|
53 |
### Training results
|
54 |
|
55 |
-
| Training Loss | Epoch | Step | Validation Loss | Wer
|
56 |
-
|
57 |
-
|
|
58 |
-
|
|
59 |
-
|
|
60 |
-
|
|
61 |
-
| 0.
|
62 |
-
| 0.
|
63 |
-
| 0.
|
64 |
-
| 0.
|
65 |
-
| 0.
|
66 |
-
| 0.
|
67 |
-
| 0.
|
68 |
-
| 0.
|
69 |
-
| 0.
|
70 |
-
| 0.4485 | 35.9 | 3500 | 0.9760 | 0.5949 | 0.2175 |
|
71 |
-
| 0.4219 | 38.46 | 3750 | 0.9824 | 0.5926 | 0.2177 |
|
72 |
-
| 0.397 | 41.03 | 4000 | 0.9669 | 0.5885 | 0.2138 |
|
73 |
-
| 0.3912 | 43.59 | 4250 | 0.9857 | 0.5908 | 0.2145 |
|
74 |
-
| 0.3764 | 46.15 | 4500 | 0.9937 | 0.5886 | 0.2145 |
|
75 |
-
| 0.3742 | 48.72 | 4750 | 0.9840 | 0.5852 | 0.2130 |
|
76 |
|
77 |
|
78 |
### Framework versions
|
79 |
|
80 |
-
- Transformers 4.
|
81 |
-
- Pytorch 2.1
|
82 |
-
- Datasets 2.
|
83 |
-
- Tokenizers 0.15.
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
+
base_model: facebook/wav2vec2-large-xlsr-53
|
4 |
tags:
|
5 |
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- common_voice_15_0
|
8 |
metrics:
|
9 |
- wer
|
10 |
model-index:
|
11 |
+
- name: wav2vec2-large-xlsr-53-br
|
12 |
+
results:
|
13 |
+
- task:
|
14 |
+
name: Automatic Speech Recognition
|
15 |
+
type: automatic-speech-recognition
|
16 |
+
dataset:
|
17 |
+
name: common_voice_15_0
|
18 |
+
type: common_voice_15_0
|
19 |
+
config: br
|
20 |
+
split: None
|
21 |
+
args: br
|
22 |
+
metrics:
|
23 |
+
- name: Wer
|
24 |
+
type: wer
|
25 |
+
value: 54.71511888739345
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
29 |
should probably proofread and complete it, then remove this comment. -->
|
30 |
|
31 |
+
# wav2vec2-large-xlsr-53-br
|
32 |
|
33 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the common_voice_15_0 dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
+
- Loss: 0.7879
|
36 |
+
- Wer: 54.7151
|
37 |
+
- Cer: 19.2493
|
38 |
|
39 |
## Model description
|
40 |
|
|
|
54 |
|
55 |
The following hyperparameters were used during training:
|
56 |
- learning_rate: 6e-05
|
57 |
+
- train_batch_size: 8
|
58 |
- eval_batch_size: 8
|
59 |
- seed: 42
|
60 |
+
- gradient_accumulation_steps: 4
|
61 |
- total_train_batch_size: 32
|
62 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
63 |
- lr_scheduler_type: linear
|
64 |
+
- lr_scheduler_warmup_steps: 300
|
65 |
+
- num_epochs: 30
|
66 |
- mixed_precision_training: Native AMP
|
67 |
|
68 |
### Training results
|
69 |
|
70 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|
71 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
|
72 |
+
| 6.3257 | 2.18 | 500 | 3.0700 | 100.0 | 99.0871 |
|
73 |
+
| 2.2071 | 4.36 | 1000 | 1.1541 | 80.0449 | 29.4230 |
|
74 |
+
| 1.0019 | 6.54 | 1500 | 0.8986 | 69.2059 | 24.3938 |
|
75 |
+
| 0.7796 | 8.71 | 2000 | 0.8015 | 63.3737 | 22.1296 |
|
76 |
+
| 0.6677 | 10.89 | 2500 | 0.8014 | 61.4984 | 21.4568 |
|
77 |
+
| 0.5937 | 13.07 | 3000 | 0.7623 | 58.9323 | 20.4929 |
|
78 |
+
| 0.5454 | 15.25 | 3500 | 0.7975 | 57.8466 | 20.2585 |
|
79 |
+
| 0.5075 | 17.43 | 4000 | 0.7831 | 56.7250 | 19.7879 |
|
80 |
+
| 0.4837 | 19.61 | 4500 | 0.7902 | 55.9623 | 19.5101 |
|
81 |
+
| 0.4529 | 21.79 | 5000 | 0.7851 | 54.9753 | 19.0924 |
|
82 |
+
| 0.4381 | 23.97 | 5500 | 0.7865 | 55.1727 | 19.3211 |
|
83 |
+
| 0.4208 | 26.14 | 6000 | 0.8168 | 55.1817 | 19.3967 |
|
84 |
+
| 0.4197 | 28.32 | 6500 | 0.7879 | 54.7151 | 19.2493 |
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
|
86 |
|
87 |
### Framework versions
|
88 |
|
89 |
+
- Transformers 4.39.1
|
90 |
+
- Pytorch 2.0.1+cu117
|
91 |
+
- Datasets 2.18.0
|
92 |
+
- Tokenizers 0.15.2
|
model.safetensors
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1262012432
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1c604794129ec1687d5ee04edcae82b445f864b30c3be119899c787a85c59eb4
|
3 |
size 1262012432
|
runs/Jun03_08-47-04_gweltaz-NUC10i7FNK/events.out.tfevents.1717397345.gweltaz-NUC10i7FNK.2937.0
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:19812879513e1a20a3853cea6e6fd5e63ad5f4132b9ec49c5c0c553a77391644
|
3 |
+
size 14168
|