sanchit-gandhi HF staff commited on
Commit
4f7b604
1 Parent(s): 6ebdee7

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +83 -0
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - es
4
+ license: apache-2.0
5
+ tags:
6
+ - hf-asr-leaderboard
7
+ - generated_from_trainer
8
+ datasets:
9
+ - facebook/multilingual_librispeech
10
+ metrics:
11
+ - wer
12
+ model-index:
13
+ - name: Whisper Small Es - Sanchit Gandhi
14
+ results:
15
+ - task:
16
+ name: Automatic Speech Recognition
17
+ type: automatic-speech-recognition
18
+ dataset:
19
+ name: Multilingual LibriSpeech
20
+ type: facebook/multilingual_librispeech
21
+ args: 'config: es, split: test'
22
+ metrics:
23
+ - name: Wer
24
+ type: wer
25
+ value: 4.426038712301834
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
+ # Whisper Small Es - Sanchit Gandhi
32
+
33
+ This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Multilingual LibriSpeech dataset.
34
+ It achieves the following results on the evaluation set:
35
+ - Loss: 0.1107
36
+ - Wer: 4.4260
37
+
38
+ ## Model description
39
+
40
+ More information needed
41
+
42
+ ## Intended uses & limitations
43
+
44
+ More information needed
45
+
46
+ ## Training and evaluation data
47
+
48
+ More information needed
49
+
50
+ ## Training procedure
51
+
52
+ ### Training hyperparameters
53
+
54
+ The following hyperparameters were used during training:
55
+ - learning_rate: 1e-05
56
+ - train_batch_size: 2
57
+ - eval_batch_size: 4
58
+ - seed: 42
59
+ - gradient_accumulation_steps: 16
60
+ - total_train_batch_size: 32
61
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
62
+ - lr_scheduler_type: linear
63
+ - lr_scheduler_warmup_steps: 500
64
+ - training_steps: 2500
65
+ - mixed_precision_training: Native AMP
66
+
67
+ ### Training results
68
+
69
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
70
+ |:-------------:|:-----:|:----:|:---------------:|:------:|
71
+ | 0.186 | 0.2 | 500 | 0.1487 | 6.1786 |
72
+ | 0.1947 | 0.4 | 1000 | 0.1350 | 5.5910 |
73
+ | 0.3566 | 0.6 | 1500 | 0.1242 | 4.9537 |
74
+ | 0.1237 | 0.8 | 2000 | 0.1181 | 4.8001 |
75
+ | 0.1902 | 1.0 | 2500 | 0.1107 | 4.4260 |
76
+
77
+
78
+ ### Framework versions
79
+
80
+ - Transformers 4.25.0.dev0
81
+ - Pytorch 1.12.0
82
+ - Datasets 2.6.2.dev0
83
+ - Tokenizers 0.12.1