Downloads · 30 days
8
11% of all-time downloads
hindi_fb1mms_timebalancedreg
odelz/hindi_fb1mms_timebalancedreg
hindi_fb1mms_timebalancedreg is a automatic speech recognition model from odelz. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as cc-by-nc-4.0.
should probably proofread and complete it, then remove this comment. --
All-time downloads
70
Public
Parameters
965M
162 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors3.9 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- License
- cc-by-nc-4.0
- Model type
- wav2vec2
- Access
- Public
- Created
- Jun 23, 2024
- Updated
- Jul 1, 2024
- SHA
- 2c10f34b
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. -->
hindi_fb1mms_timebalancedreg
This model is a fine-tuned version of facebook/mms-1b-all on the audiofolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.7182
- Wer: 0.4259
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: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 4.087 | 1.0191 | 400 | 3.5884 | 0.9998 |
| 3.935 | 2.0382 | 800 | 3.4190 | 0.9959 |
| 3.3712 | 3.0573 | 1200 | 3.3003 | 0.9709 |
| 3.2027 | 4.0764 | 1600 | 2.8687 | 0.9861 |
| 1.4667 | 5.0955 | 2000 | 0.6547 | 0.4129 |
| 1.2468 | 6.1146 | 2400 | 0.6031 | 0.3955 |
| 1.2401 | 7.1338 | 2800 | 0.6334 | 0.4172 |
| 1.2952 | 8.1529 | 3200 | 0.6857 | 0.4238 |
| 1.2466 | 9.1720 | 3600 | 0.7279 | 0.4361 |
| 1.2094 | 10.1911 | 4000 | 0.6768 | 0.4140 |
| 1.1764 | 11.2102 | 4400 | 0.6735 | 0.4234 |
| 1.1491 | 12.2293 | 4800 | 0.7047 | 0.4334 |
| 1.1504 | 13.2484 | 5200 | 0.6704 | 0.4215 |
| 1.1656 | 14.2675 | 5600 | 0.6684 | 0.4207 |
| 1.1666 | 15.2866 | 6000 | 0.7367 | 0.4339 |
| 1.1512 | 16.3057 | 6400 | 0.7384 | 0.4386 |
| 1.1646 | 17.3248 | 6800 | 0.7087 | 0.4251 |
| 1.1407 | 18.3439 | 7200 | 0.7192 | 0.4329 |
| 1.1207 | 19.3631 | 7600 | 0.7141 | 0.4236 |
| 1.1145 | 20.3822 | 8000 | 0.7503 | 0.4374 |
| 1.1138 | 21.4013 | 8400 | 0.7235 | 0.4278 |
| 1.1091 | 22.4204 | 8800 | 0.7468 | 0.4404 |
| 1.1255 | 23.4395 | 9200 | 0.7177 | 0.4264 |
| 1.0959 | 24.4586 | 9600 | 0.7050 | 0.4191 |
| 1.106 | 25.4777 | 10000 | 0.7420 | 0.4337 |
| 1.0949 | 26.4968 | 10400 | 0.7063 | 0.4223 |
| 1.1142 | 27.5159 | 10800 | 0.7170 | 0.4257 |
| 1.1076 | 28.5350 | 11200 | 0.7223 | 0.4267 |
| 1.1028 | 29.5541 | 11600 | 0.7182 | 0.4259 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.1+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- wav2vec2
- License
- cc-by-nc-4.0
- Created
- Jun 23, 2024
- Updated
- Jul 1, 2024