File size: 1,629 Bytes
6cd73fa |
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 |
---
language:
- zh
base_model: ZhihCheng/whisper-base-zh
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper base ZH
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 base ZH
This model is a fine-tuned version of [ZhihCheng/whisper-base-zh](https://huggingface.co/ZhihCheng/whisper-base-zh) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2200
- Wer: 44.8955
## 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: 16
- 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: 4000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.1704 | 0.27 | 1000 | 0.2787 | 51.6077 |
| 0.1479 | 0.54 | 2000 | 0.2480 | 48.4124 |
| 0.1145 | 0.81 | 3000 | 0.2289 | 46.1415 |
| 0.076 | 1.08 | 4000 | 0.2200 | 44.8955 |
### Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.2.0a0+6a974be
- Datasets 2.16.1
- Tokenizers 0.15.1
|