update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- generated_from_trainer
|
| 4 |
+
metrics:
|
| 5 |
+
- wer
|
| 6 |
+
model-index:
|
| 7 |
+
- name: dgx2_whisper_small_mozilla_noisy_distil_epochs_50_batch_8
|
| 8 |
+
results: []
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 12 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 13 |
+
|
| 14 |
+
# dgx2_whisper_small_mozilla_noisy_distil_epochs_50_batch_8
|
| 15 |
+
|
| 16 |
+
This model is a fine-tuned version of [rohitp1/kkkh_whisper_small_distillation_att_loss_mozilla_epochs_100_batch_4_concat_dataset](https://huggingface.co/rohitp1/kkkh_whisper_small_distillation_att_loss_mozilla_epochs_100_batch_4_concat_dataset) on the None dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 1.1254
|
| 19 |
+
- Wer: 20.5209
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-05
|
| 39 |
+
- train_batch_size: 8
|
| 40 |
+
- eval_batch_size: 1
|
| 41 |
+
- seed: 42
|
| 42 |
+
- gradient_accumulation_steps: 256
|
| 43 |
+
- total_train_batch_size: 2048
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: cosine_with_restarts
|
| 46 |
+
- lr_scheduler_warmup_ratio: 0.2
|
| 47 |
+
- num_epochs: 50
|
| 48 |
+
- mixed_precision_training: Native AMP
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 53 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|
|
| 54 |
+
| 0.0006 | 4.39 | 150 | 0.5538 | 19.3893 |
|
| 55 |
+
| 0.1453 | 8.78 | 300 | 0.7751 | 20.7263 |
|
| 56 |
+
| 0.3233 | 13.17 | 450 | 0.8857 | 20.7994 |
|
| 57 |
+
| 0.486 | 17.55 | 600 | 1.0980 | 20.6462 |
|
| 58 |
+
| 0.6433 | 21.94 | 750 | 1.1264 | 20.5835 |
|
| 59 |
+
| 0.6452 | 26.33 | 900 | 1.1254 | 20.5209 |
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
### Framework versions
|
| 63 |
+
|
| 64 |
+
- Transformers 4.28.1
|
| 65 |
+
- Pytorch 1.12.1
|
| 66 |
+
- Datasets 2.8.0
|
| 67 |
+
- Tokenizers 0.13.2
|