Hugging Face's logo

CLEAR-Global
/
w2v-bert-2.0-chichewa_34_136h

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2-bert
CLEAR-Global/chichewa_34_136h
Generated from Trainer
Model card Files Metrics Community
ayymen commited on
Commit
f9d74d1
·
verified ·
1 Parent(s): 4999634

Model save

Browse files
Files changed (2) hide show
  1. README.md +83 -0
  2. model.safetensors +1 -1
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ license: mit
4
+ base_model: facebook/w2v-bert-2.0
5
+ tags:
6
+ - generated_from_trainer
7
+ metrics:
8
+ - wer
9
+ model-index:
10
+ - name: w2v-bert-2.0-chichewa_34_136h
11
+ results: []
12
+ ---
13
+
14
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
15
+ should probably proofread and complete it, then remove this comment. -->
16
+
17
+ # w2v-bert-2.0-chichewa_34_136h
18
+
19
+ This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on an unknown dataset.
20
+ It achieves the following results on the evaluation set:
21
+ - Loss: 0.3207
22
+ - Wer: 0.3914
23
+ - Cer: 0.1153
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 3e-05
43
+ - train_batch_size: 32
44
+ - eval_batch_size: 32
45
+ - seed: 42
46
+ - gradient_accumulation_steps: 2
47
+ - total_train_batch_size: 64
48
+ - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
49
+ - lr_scheduler_type: linear
50
+ - lr_scheduler_warmup_ratio: 0.1
51
+ - training_steps: 100000
52
+ - mixed_precision_training: Native AMP
53
+
54
+ ### Training results
55
+
56
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
57
+ |:-------------:|:-------:|:-----:|:---------------:|:------:|:------:|
58
+ | 2.7429 | 0.6122 | 1000 | 2.9154 | 0.9860 | 0.8820 |
59
+ | 0.1586 | 1.2241 | 2000 | 0.7989 | 0.6341 | 0.1888 |
60
+ | 0.0475 | 1.8362 | 3000 | 0.7777 | 0.5725 | 0.1637 |
61
+ | 0.0452 | 2.4481 | 4000 | 0.4482 | 0.5083 | 0.1482 |
62
+ | 0.0387 | 3.0600 | 5000 | 0.4168 | 0.4770 | 0.1396 |
63
+ | 0.0454 | 3.6722 | 6000 | 0.3792 | 0.4501 | 0.1306 |
64
+ | 0.0215 | 4.2841 | 7000 | 0.3758 | 0.4564 | 0.1324 |
65
+ | 0.0342 | 4.8962 | 8000 | 0.3737 | 0.4557 | 0.1298 |
66
+ | 0.0243 | 5.5081 | 9000 | 0.3805 | 0.4325 | 0.1252 |
67
+ | 0.0183 | 6.1200 | 10000 | 0.3490 | 0.4257 | 0.1240 |
68
+ | 0.0253 | 6.7322 | 11000 | 0.3670 | 0.4185 | 0.1199 |
69
+ | 0.0115 | 7.3440 | 12000 | 0.3664 | 0.4125 | 0.1207 |
70
+ | 0.0141 | 7.9562 | 13000 | 0.2952 | 0.4021 | 0.1153 |
71
+ | 0.0141 | 8.5681 | 14000 | 0.3231 | 0.4031 | 0.1133 |
72
+ | 0.0082 | 9.1800 | 15000 | 0.3209 | 0.4000 | 0.1141 |
73
+ | 0.0214 | 9.7922 | 16000 | 0.3115 | 0.3985 | 0.1134 |
74
+ | 0.0146 | 10.4040 | 17000 | 0.3092 | 0.3743 | 0.1089 |
75
+ | 0.0367 | 11.0159 | 18000 | 0.3207 | 0.3914 | 0.1153 |
76
+
77
+
78
+ ### Framework versions
79
+
80
+ - Transformers 4.48.1
81
+ - Pytorch 2.6.0+cu124
82
+ - Datasets 3.5.0
83
+ - Tokenizers 0.21.1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9996d4ae1b6eb777ffef504d92df7d476195ecacc1670bec9ab99835d090323a
3
  size 2423068760
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5d7706daddab36d529263ce0c78ebe86559e96099c5d730aad82dde3a6c74eb
3
  size 2423068760