File size: 1,632 Bytes
8e8bd44 1556c6b 8e8bd44 2fe25c9 8e8bd44 2fe25c9 8e8bd44 2fe25c9 8e8bd44 2fe25c9 8e8bd44 2fe25c9 8e8bd44 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 |
---
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-tiny-hi2_test
results: []
---
<!-- 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-tiny-hi2_test
This model was trained from scratch on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4940
- Wer: 59.7206
## 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: 3.75e-05
- train_batch_size: 16
- eval_batch_size: 4
- 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: constant
- lr_scheduler_warmup_steps: 50
- training_steps: 200
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:------:|:----:|:---------------:|:-------:|
| 0.6766 | 1.2698 | 40 | 0.6154 | 81.4733 |
| 0.3599 | 2.5397 | 80 | 0.5078 | 67.0110 |
| 0.2297 | 3.8095 | 120 | 0.4940 | 59.7206 |
| 0.153 | 5.0794 | 160 | 0.5193 | 62.0745 |
| 0.0885 | 6.3492 | 200 | 0.5557 | 60.5843 |
### Framework versions
- Transformers 4.41.1
- Pytorch 2.3.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
|