Downloads · 30 days
8
29% of all-time downloads
whisper-small-ko-E2
Gummybear05/whisper-small-ko-E2
whisper-small-ko-E2 is a automatic speech recognition model from Gummybear05. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as apache-2.0.
should probably proofread and complete it, then remove this comment. --
All-time downloads
28
Public
Parameters
242M
30 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors967 MB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- License
- apache-2.0
- Model type
- whisper
- Access
- Public
- Created
- Dec 12, 2023
- Updated
- Dec 12, 2023
- SHA
- 6a24668f
Base models
Datasets
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
whisper-small-ko-E2
This model is a fine-tuned version of openai/whisper-small on the aihub elder over 70 dataset. It achieves the following results on the evaluation set:
- Loss: 0.1664
- Cer: 4.4232
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.319 | 0.13 | 100 | 0.2515 | 6.9431 |
| 0.2959 | 0.26 | 200 | 0.2258 | 6.1501 |
| 0.3025 | 0.39 | 300 | 0.2085 | 5.5275 |
| 0.2603 | 0.52 | 400 | 0.2091 | 5.5334 |
| 0.2436 | 0.64 | 500 | 0.1952 | 5.3454 |
| 0.2213 | 0.77 | 600 | 0.1903 | 5.0282 |
| 0.2142 | 0.9 | 700 | 0.1836 | 5.0811 |
| 0.1257 | 1.03 | 800 | 0.1783 | 6.0973 |
| 0.0992 | 1.16 | 900 | 0.1745 | 4.6640 |
| 0.1092 | 1.29 | 1000 | 0.1832 | 5.7390 |
| 0.1134 | 1.42 | 1100 | 0.1754 | 5.0458 |
| 0.093 | 1.55 | 1200 | 0.1731 | 4.5348 |
| 0.1194 | 1.68 | 1300 | 0.1671 | 4.2822 |
| 0.1016 | 1.81 | 1400 | 0.1687 | 4.3997 |
| 0.1002 | 1.93 | 1500 | 0.1759 | 6.1031 |
| 0.0374 | 2.06 | 1600 | 0.1702 | 4.6464 |
| 0.0404 | 2.19 | 1700 | 0.1695 | 4.8696 |
| 0.0374 | 2.32 | 1800 | 0.1712 | 4.5230 |
| 0.039 | 2.45 | 1900 | 0.1673 | 4.2234 |
| 0.0404 | 2.58 | 2000 | 0.1671 | 4.5406 |
| 0.0422 | 2.71 | 2100 | 0.1653 | 4.3057 |
| 0.032 | 2.84 | 2200 | 0.1683 | 4.8755 |
| 0.0329 | 2.97 | 2300 | 0.1657 | 3.9944 |
| 0.0119 | 3.09 | 2400 | 0.1687 | 4.0942 |
| 0.0143 | 3.22 | 2500 | 0.1676 | 4.2704 |
| 0.0119 | 3.35 | 2600 | 0.1687 | 4.2822 |
| 0.011 | 3.48 | 2700 | 0.1714 | 4.2939 |
| 0.0111 | 3.61 | 2800 | 0.1657 | 4.3527 |
| 0.0124 | 3.74 | 2900 | 0.1661 | 4.2058 |
| 0.0123 | 3.87 | 3000 | 0.1662 | 4.4114 |
| 0.0127 | 4.0 | 3100 | 0.1664 | 4.4232 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- whisper
- License
- apache-2.0
- Languages
- hi
- Created
- Dec 12, 2023
- Updated
- Dec 12, 2023