p6_commonvoice_hc / README.md
honzapucalek's picture
Model save
7b50e4d verified
|
raw
history blame
2.05 kB
---
license: apache-2.0
base_model: honzapucalek/p6_commonvoice_16_1
tags:
- generated_from_trainer
datasets:
- p6_hc
metrics:
- wer
model-index:
- name: p6_commonvoice_hc
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: p6_hc
type: p6_hc
config: cs
split: test
args: cs
metrics:
- name: Wer
type: wer
value: 0.17960769800148038
---
<!-- 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. -->
# p6_commonvoice_hc
This model is a fine-tuned version of [honzapucalek/p6_commonvoice_16_1](https://huggingface.co./honzapucalek/p6_commonvoice_16_1) on the p6_hc dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5960
- Wer: 0.1796
## 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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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 | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.0024 | 14.49 | 1000 | 0.4506 | 0.2335 |
| 0.0024 | 28.99 | 2000 | 0.4568 | 0.1868 |
| 0.0001 | 43.48 | 3000 | 0.5552 | 0.1823 |
| 0.0001 | 57.97 | 4000 | 0.5876 | 0.1807 |
| 0.0001 | 72.46 | 5000 | 0.5960 | 0.1796 |
### Framework versions
- Transformers 4.37.2
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1