File size: 1,762 Bytes
c59468b |
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 68 |
---
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper_finetuned_ver2
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_finetuned_ver2
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0048
- Cer: 0.5262
- Wer: 0.4840
## 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: 1e-05
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|
| 0.0 | 35.71 | 1000 | 0.0047 | 0.5496 | 0.5227 |
| 0.0001 | 71.43 | 2000 | 0.0048 | 0.5262 | 0.4840 |
| 0.0 | 107.14 | 3000 | 0.0051 | 0.5964 | 0.5615 |
| 0.0 | 142.86 | 4000 | 0.0053 | 0.6080 | 0.5808 |
| 0.0 | 178.57 | 5000 | 0.0054 | 0.6080 | 0.5808 |
### Framework versions
- Transformers 4.40.0.dev0
- Pytorch 2.0.1
- Datasets 2.18.0
- Tokenizers 0.15.2
|