Commit
•
67e001c
1
Parent(s):
e77d60b
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- as
|
4 |
+
license: apache-2.0
|
5 |
+
tags:
|
6 |
+
- whisper-event
|
7 |
+
- generated_from_trainer
|
8 |
+
datasets:
|
9 |
+
- mozilla-foundation/common_voice_11_0
|
10 |
+
metrics:
|
11 |
+
- wer
|
12 |
+
model-index:
|
13 |
+
- name: Whisper Small Assamese - Drishti Sharma
|
14 |
+
results:
|
15 |
+
- task:
|
16 |
+
name: Automatic Speech Recognition
|
17 |
+
type: automatic-speech-recognition
|
18 |
+
dataset:
|
19 |
+
name: Common Voice 11.0
|
20 |
+
type: mozilla-foundation/common_voice_11_0
|
21 |
+
config: as
|
22 |
+
split: test
|
23 |
+
args: as
|
24 |
+
metrics:
|
25 |
+
- name: Wer
|
26 |
+
type: wer
|
27 |
+
value: 30.650219527493206
|
28 |
+
---
|
29 |
+
|
30 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
31 |
+
should probably proofread and complete it, then remove this comment. -->
|
32 |
+
|
33 |
+
# Whisper Small Assamese - Drishti Sharma
|
34 |
+
|
35 |
+
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
|
36 |
+
It achieves the following results on the evaluation set:
|
37 |
+
- Loss: 0.5164
|
38 |
+
- Wer: 30.6502
|
39 |
+
|
40 |
+
## Model description
|
41 |
+
|
42 |
+
More information needed
|
43 |
+
|
44 |
+
## Intended uses & limitations
|
45 |
+
|
46 |
+
More information needed
|
47 |
+
|
48 |
+
## Training and evaluation data
|
49 |
+
|
50 |
+
More information needed
|
51 |
+
|
52 |
+
## Training procedure
|
53 |
+
|
54 |
+
### Training hyperparameters
|
55 |
+
|
56 |
+
The following hyperparameters were used during training:
|
57 |
+
- learning_rate: 1e-05
|
58 |
+
- train_batch_size: 16
|
59 |
+
- eval_batch_size: 8
|
60 |
+
- seed: 42
|
61 |
+
- gradient_accumulation_steps: 4
|
62 |
+
- total_train_batch_size: 64
|
63 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
64 |
+
- lr_scheduler_type: linear
|
65 |
+
- lr_scheduler_warmup_steps: 150
|
66 |
+
- training_steps: 500
|
67 |
+
- mixed_precision_training: Native AMP
|
68 |
+
|
69 |
+
### Training results
|
70 |
+
|
71 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
72 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|
|
73 |
+
| 1.1539 | 3.02 | 50 | 0.7835 | 89.0863 |
|
74 |
+
| 0.2089 | 7.0 | 100 | 0.3041 | 37.7378 |
|
75 |
+
| 0.0428 | 10.02 | 150 | 0.3760 | 33.9118 |
|
76 |
+
| 0.0141 | 14.01 | 200 | 0.4400 | 31.6538 |
|
77 |
+
| 0.0059 | 17.02 | 250 | 0.4472 | 31.2774 |
|
78 |
+
| 0.0022 | 21.01 | 300 | 0.4696 | 31.0475 |
|
79 |
+
| 0.0005 | 24.03 | 350 | 0.5032 | 31.2983 |
|
80 |
+
| 0.0003 | 28.02 | 400 | 0.5051 | 30.7129 |
|
81 |
+
| 0.0003 | 32.0 | 450 | 0.5137 | 30.7338 |
|
82 |
+
| 0.0003 | 35.02 | 500 | 0.5164 | 30.6502 |
|
83 |
+
|
84 |
+
|
85 |
+
### Framework versions
|
86 |
+
|
87 |
+
- Transformers 4.26.0.dev0
|
88 |
+
- Pytorch 1.13.0+cu116
|
89 |
+
- Datasets 2.7.1.dev0
|
90 |
+
- Tokenizers 0.13.2
|