update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- common_voice
|
| 7 |
+
model-index:
|
| 8 |
+
- name: xls-r-300m-yaswanth-hindi2
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# xls-r-300m-yaswanth-hindi2
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 1.7163
|
| 20 |
+
- Wer: 0.6951
|
| 21 |
+
|
| 22 |
+
## Model description
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Intended uses & limitations
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training and evaluation data
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training procedure
|
| 35 |
+
|
| 36 |
+
### Training hyperparameters
|
| 37 |
+
|
| 38 |
+
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 0.0007
|
| 40 |
+
- train_batch_size: 32
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 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 |
+
- num_epochs: 100
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 53 |
+
| 4.986 | 4.46 | 500 | 2.0194 | 1.1857 |
|
| 54 |
+
| 0.9232 | 8.93 | 1000 | 1.2665 | 0.8435 |
|
| 55 |
+
| 0.5094 | 13.39 | 1500 | 1.2473 | 0.7893 |
|
| 56 |
+
| 0.3618 | 17.86 | 2000 | 1.3675 | 0.7789 |
|
| 57 |
+
| 0.2914 | 22.32 | 2500 | 1.3725 | 0.7914 |
|
| 58 |
+
| 0.2462 | 26.79 | 3000 | 1.4567 | 0.7795 |
|
| 59 |
+
| 0.228 | 31.25 | 3500 | 1.6179 | 0.7872 |
|
| 60 |
+
| 0.1995 | 35.71 | 4000 | 1.4932 | 0.7555 |
|
| 61 |
+
| 0.1878 | 40.18 | 4500 | 1.5352 | 0.7480 |
|
| 62 |
+
| 0.165 | 44.64 | 5000 | 1.5238 | 0.7440 |
|
| 63 |
+
| 0.1514 | 49.11 | 5500 | 1.5842 | 0.7498 |
|
| 64 |
+
| 0.1416 | 53.57 | 6000 | 1.6662 | 0.7524 |
|
| 65 |
+
| 0.1351 | 58.04 | 6500 | 1.6280 | 0.7356 |
|
| 66 |
+
| 0.1196 | 62.5 | 7000 | 1.6329 | 0.7250 |
|
| 67 |
+
| 0.1109 | 66.96 | 7500 | 1.6435 | 0.7302 |
|
| 68 |
+
| 0.1008 | 71.43 | 8000 | 1.7058 | 0.7170 |
|
| 69 |
+
| 0.0907 | 75.89 | 8500 | 1.6880 | 0.7387 |
|
| 70 |
+
| 0.0816 | 80.36 | 9000 | 1.6957 | 0.7031 |
|
| 71 |
+
| 0.0743 | 84.82 | 9500 | 1.7547 | 0.7222 |
|
| 72 |
+
| 0.0694 | 89.29 | 10000 | 1.6974 | 0.7117 |
|
| 73 |
+
| 0.0612 | 93.75 | 10500 | 1.7251 | 0.7020 |
|
| 74 |
+
| 0.0577 | 98.21 | 11000 | 1.7163 | 0.6951 |
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
### Framework versions
|
| 78 |
+
|
| 79 |
+
- Transformers 4.16.0
|
| 80 |
+
- Pytorch 1.10.0+cu111
|
| 81 |
+
- Datasets 1.18.3
|
| 82 |
+
- Tokenizers 0.11.0
|