update model card README.md
Browse files
README.md
CHANGED
|
@@ -7,9 +7,23 @@ tags:
|
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- mozilla-foundation/common_voice_11_0
|
|
|
|
|
|
|
| 10 |
model-index:
|
| 11 |
- name: Whisper Small Ckb - Razhan Hameed
|
| 12 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
---
|
| 14 |
|
| 15 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -18,6 +32,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
# Whisper Small Ckb - Razhan Hameed
|
| 19 |
|
| 20 |
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -43,9 +60,27 @@ The following hyperparameters were used during training:
|
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_steps: 500
|
| 46 |
-
- training_steps:
|
| 47 |
- mixed_precision_training: Native AMP
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
### Framework versions
|
| 50 |
|
| 51 |
- Transformers 4.26.0.dev0
|
|
|
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- mozilla-foundation/common_voice_11_0
|
| 10 |
+
metrics:
|
| 11 |
+
- wer
|
| 12 |
model-index:
|
| 13 |
- name: Whisper Small Ckb - Razhan Hameed
|
| 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: null
|
| 22 |
+
split: test
|
| 23 |
+
metrics:
|
| 24 |
+
- name: Wer
|
| 25 |
+
type: wer
|
| 26 |
+
value: 33.2192952446117
|
| 27 |
---
|
| 28 |
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 32 |
# Whisper Small Ckb - Razhan Hameed
|
| 33 |
|
| 34 |
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
|
| 35 |
+
It achieves the following results on the evaluation set:
|
| 36 |
+
- Loss: 0.3825
|
| 37 |
+
- Wer: 33.2193
|
| 38 |
|
| 39 |
## Model description
|
| 40 |
|
|
|
|
| 60 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 61 |
- lr_scheduler_type: linear
|
| 62 |
- lr_scheduler_warmup_steps: 500
|
| 63 |
+
- training_steps: 12000
|
| 64 |
- mixed_precision_training: Native AMP
|
| 65 |
|
| 66 |
+
### Training results
|
| 67 |
+
|
| 68 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 69 |
+
|:-------------:|:-----:|:-----:|:---------------:|:-------:|
|
| 70 |
+
| 0.1693 | 2.49 | 1000 | 0.2060 | 39.1265 |
|
| 71 |
+
| 0.0722 | 4.98 | 2000 | 0.2124 | 36.3173 |
|
| 72 |
+
| 0.0127 | 7.46 | 3000 | 0.2736 | 36.5568 |
|
| 73 |
+
| 0.008 | 9.95 | 4000 | 0.3131 | 35.7015 |
|
| 74 |
+
| 0.0032 | 12.44 | 5000 | 0.3434 | 35.3936 |
|
| 75 |
+
| 0.0028 | 14.93 | 6000 | 0.3453 | 35.9258 |
|
| 76 |
+
| 0.003 | 17.41 | 7000 | 0.3558 | 34.9565 |
|
| 77 |
+
| 0.0022 | 19.9 | 8000 | 0.3593 | 34.2722 |
|
| 78 |
+
| 0.0016 | 22.39 | 9000 | 0.3639 | 34.3369 |
|
| 79 |
+
| 0.0015 | 24.88 | 10000 | 0.3785 | 34.0062 |
|
| 80 |
+
| 0.0009 | 27.36 | 11000 | 0.3915 | 34.2951 |
|
| 81 |
+
| 0.0001 | 29.85 | 12000 | 0.3825 | 33.2193 |
|
| 82 |
+
|
| 83 |
+
|
| 84 |
### Framework versions
|
| 85 |
|
| 86 |
- Transformers 4.26.0.dev0
|