Hugging Face's logo

makhataei
/
Whisper-Small-Common-Voice

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Persian
whisper
fa-asr
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
makhataei commited on
Commit
8df4d03
·
verified ·
1 Parent(s): ca334d1

End of training

Browse files
Files changed (3) hide show
  1. README.md +21 -49
  2. model.safetensors +1 -1
  3. runs/Apr14_16-37-37_Software-AI/events.out.tfevents.1713100059.Software-AI.5448.0 +2 -2
README.md CHANGED
@@ -6,6 +6,8 @@ base_model: makhataei/Whisper-Small-Common-Voice
6
  tags:
7
  - fa-asr
8
  - generated_from_trainer
 
 
9
  metrics:
10
  - wer
11
  model-index:
@@ -18,10 +20,10 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  # Whisper Small Persian
20
 
21
- This model is a fine-tuned version of [makhataei/Whisper-Small-Common-Voice](https://huggingface.co/makhataei/Whisper-Small-Common-Voice) on the Ctejarat dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.5349
24
- - Wer: 26.2116
25
 
26
  ## Model description
27
 
@@ -44,57 +46,27 @@ The following hyperparameters were used during training:
44
  - train_batch_size: 10
45
  - eval_batch_size: 10
46
  - seed: 42
47
- - gradient_accumulation_steps: 8
48
- - total_train_batch_size: 80
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: linear
51
- - lr_scheduler_warmup_steps: 100
52
- - training_steps: 4000
53
 
54
  ### Training results
55
 
56
- | Training Loss | Epoch | Step | Validation Loss | Wer |
57
- |:-------------:|:------:|:----:|:---------------:|:-------:|
58
- | 0.2944 | 9.64 | 100 | 0.4843 | 33.6519 |
59
- | 0.1048 | 19.28 | 200 | 0.4394 | 30.1706 |
60
- | 0.0273 | 28.92 | 300 | 0.4493 | 29.7611 |
61
- | 0.0083 | 38.55 | 400 | 0.4645 | 29.4198 |
62
- | 0.0042 | 48.19 | 500 | 0.4744 | 28.5324 |
63
- | 0.0026 | 57.83 | 600 | 0.4811 | 28.3276 |
64
- | 0.0018 | 67.47 | 700 | 0.4863 | 27.6451 |
65
- | 0.0014 | 77.11 | 800 | 0.4907 | 27.7816 |
66
- | 0.0012 | 86.75 | 900 | 0.4945 | 27.4403 |
67
- | 0.0009 | 96.39 | 1000 | 0.4979 | 27.4403 |
68
- | 0.0008 | 106.02 | 1100 | 0.5010 | 26.8259 |
69
- | 0.0007 | 115.66 | 1200 | 0.5036 | 26.8259 |
70
- | 0.0006 | 125.3 | 1300 | 0.5062 | 26.6894 |
71
- | 0.0006 | 134.94 | 1400 | 0.5085 | 26.3481 |
72
- | 0.0005 | 144.58 | 1500 | 0.5107 | 26.3481 |
73
- | 0.0004 | 154.22 | 1600 | 0.5126 | 26.4164 |
74
- | 0.0004 | 163.86 | 1700 | 0.5145 | 26.4846 |
75
- | 0.0004 | 173.49 | 1800 | 0.5163 | 26.3481 |
76
- | 0.0003 | 183.13 | 1900 | 0.5179 | 30.8532 |
77
- | 0.0003 | 192.77 | 2000 | 0.5194 | 30.8532 |
78
- | 0.0003 | 202.41 | 2100 | 0.5209 | 30.7850 |
79
- | 0.0003 | 212.05 | 2200 | 0.5222 | 30.9215 |
80
- | 0.0003 | 221.69 | 2300 | 0.5236 | 30.9215 |
81
- | 0.0003 | 231.33 | 2400 | 0.5248 | 30.9215 |
82
- | 0.0002 | 240.96 | 2500 | 0.5259 | 30.9215 |
83
- | 0.0002 | 250.6 | 2600 | 0.5270 | 30.7167 |
84
- | 0.0002 | 260.24 | 2700 | 0.5280 | 30.8532 |
85
- | 0.0002 | 269.88 | 2800 | 0.5290 | 30.8532 |
86
- | 0.0002 | 279.52 | 2900 | 0.5299 | 30.7167 |
87
- | 0.0002 | 289.16 | 3000 | 0.5306 | 30.7167 |
88
- | 0.0002 | 298.8 | 3100 | 0.5314 | 30.7167 |
89
- | 0.0002 | 308.43 | 3200 | 0.5321 | 30.7167 |
90
- | 0.0002 | 318.07 | 3300 | 0.5327 | 30.7850 |
91
- | 0.0002 | 327.71 | 3400 | 0.5333 | 30.7167 |
92
- | 0.0002 | 337.35 | 3500 | 0.5337 | 30.7167 |
93
- | 0.0002 | 346.99 | 3600 | 0.5341 | 30.7167 |
94
- | 0.0002 | 356.63 | 3700 | 0.5344 | 30.6485 |
95
- | 0.0002 | 366.27 | 3800 | 0.5347 | 26.2116 |
96
- | 0.0002 | 375.9 | 3900 | 0.5348 | 26.2116 |
97
- | 0.0002 | 385.54 | 4000 | 0.5349 | 26.2116 |
98
 
99
 
100
  ### Framework versions
 
6
  tags:
7
  - fa-asr
8
  - generated_from_trainer
9
+ datasets:
10
+ - mozilla-foundation/common_voice_17_0
11
  metrics:
12
  - wer
13
  model-index:
 
20
 
21
  # Whisper Small Persian
22
 
23
+ This model is a fine-tuned version of [makhataei/Whisper-Small-Common-Voice](https://huggingface.co/makhataei/Whisper-Small-Common-Voice) on the Common Voice 17.0 dataset.
24
  It achieves the following results on the evaluation set:
25
+ - Loss: 0.7219
26
+ - Wer: 46.5235
27
 
28
  ## Model description
29
 
 
46
  - train_batch_size: 10
47
  - eval_batch_size: 10
48
  - seed: 42
49
+ - gradient_accumulation_steps: 4
50
+ - total_train_batch_size: 40
51
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
52
  - lr_scheduler_type: linear
53
+ - lr_scheduler_warmup_steps: 50
54
+ - training_steps: 1000
55
 
56
  ### Training results
57
 
58
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
59
+ |:-------------:|:-----:|:----:|:---------------:|:-------:|
60
+ | 0.0057 | 0.15 | 100 | 0.7193 | 45.7885 |
61
+ | 0.0051 | 0.31 | 200 | 0.7189 | 44.8246 |
62
+ | 0.0027 | 0.46 | 300 | 0.7138 | 46.2762 |
63
+ | 0.0025 | 0.61 | 400 | 0.7172 | 45.8786 |
64
+ | 0.0029 | 0.77 | 500 | 0.7175 | 47.2909 |
65
+ | 0.0071 | 0.92 | 600 | 0.7199 | 46.7847 |
66
+ | 0.0273 | 1.07 | 700 | 0.7144 | 46.7616 |
67
+ | 0.0015 | 1.23 | 800 | 0.7174 | 46.6391 |
68
+ | 0.0016 | 1.38 | 900 | 0.7211 | 46.4958 |
69
+ | 0.0013 | 1.53 | 1000 | 0.7219 | 46.5235 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
70
 
71
 
72
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2762e9e03383f3b5c5fda67cf425f385c42c85800fc21102ad7e6f13509ce2e0
3
  size 966995080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d94bb81ecb4b8454a6a6f7413afe469ea38a87f05caa7b189b09f9eabde6e3db
3
  size 966995080
runs/Apr14_16-37-37_Software-AI/events.out.tfevents.1713100059.Software-AI.5448.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fc285640141aa6806c0e10cc351d0a89c8ea2658b7e1523a6e09716374b5dcca
3
- size 7899
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f9e18216a36380cfa402bf81b2c9546d07a951fc9b364c4ff0df1562f537f36
3
+ size 14875