File size: 1,730 Bytes
3049690 c9a7343 3049690 c9a7343 3049690 c9a7343 3049690 c9a7343 3049690 |
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 |
---
language:
- sv
license: apache-2.0
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_11_0
model-index:
- name: whisper-small-se
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-small-se
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
## Model description
The model was initially trained on 680 000 hours of audio with corresponding transcripts from the internet, 65% of which was in english audio and 83 % of which had english transcripts.
The model was then further trained for 4000 iterations, 500 of which as warm-up, on Swedish data from [Common_voice 11.0](https://huggingface.co/datasets/mozilla-foundation/common_voice_11_0). Achieving a WER of 19.865.
## 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: 500
- mixed_precision_training: Native AMP
### Training results
## Model Plot
<details>
<summary>View Training Plots</summary>
</details>
### Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2 |