Skip to content
ModelListedExternal
Download

Whisper Small Myanmar

This model is a fine-tuned version of openai/whisper-small on the chuuhtetnaing/myanmar-speech-dataset-openslr-80 dataset. It achieves the following results on the evaluation set: Loss: 0.1904 Wer: 49.0650

Sourcehuggingface.co/axtr05/whisper-small-myanmar/tree/f2118484c5abc2fb6c4fde2749e54efad573b574
KindText modelBase modelopenai/whisper-smallVersionFP32Bgrade

Model source

Kind
Text model
File size
922 MB
Base model
openai/whisper-small
Version
vFP32
Updated
Oct 31, 2026
Downloads
15
Source
Hugging Face

Source description

whisper-small-myanmar

This model is a fine-tuned version of openai/whisper-small on the chuuhtetnaing/myanmar-speech-dataset-openslr-80 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1904
  • Wer: 49.0650

Usage

from datasets import Audio, load_dataset
from transformers import pipeline

# Load a sample audio
dataset = load_dataset("chuuhtetnaing/myanmar-speech-dataset-openslr-80")
dataset = dataset.cast_column("audio", Audio(sampling_rate=16000))
test_dataset = dataset['test']
input_speech = test_dataset[42]['audio']

pipe = pipeline(model='chuuhtetnaing/whisper-small-myanmar')

output = pipe(input_speech, generate_kwargs={"language": "myanmar", "task": "transcribe"})
print(output['text']) # ကျွန်မ ပြည်ပ မှာ ပညာသင် တော့ စာမေးပွဲ ကို တပတ်တခါ စစ်တယ်

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0003
  • train_batch_size: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 200
  • num_epochs: 30
  • mixed_precision_training: Native AMP

Training results

Training LossEpochStepValidation LossWer
1.25661.0360.8893215.0045
0.88622.0720.6243388.6465
0.35463.01080.2046316.8744
0.18394.01440.169581.3001
0.11985.01800.138563.8914
0.09696.02160.158366.0285
0.0847.02520.153970.6589
0.06288.02880.160361.3090
0.05659.03240.142460.3295
0.035510.03600.145758.1478
0.029911.03960.154757.7916
0.018312.04320.154354.3633
0.013113.04680.153254.1407
0.01114.05040.160453.8736
0.008315.05400.163054.0516
0.004216.05760.171152.1371
0.003417.06120.167052.5824
0.002218.06480.164952.5378
0.001319.06840.180252.1817
0.001420.07200.182053.1612
0.00221.07560.179252.7159
0.001622.07920.179650.7124
0.000423.08280.180350.4007
0.000324.08640.180449.4657
0.000125.09000.181949.2431
0.026.09360.185749.0205
0.027.09720.187949.1541
0.028.010080.189349.1095
0.029.010440.190149.1095
0.030.010800.190449.0650

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.1+cu121
  • Datasets 2.14.5
  • Tokenizers 0.15.1

Sources

1 source
Hugging FacePrimaryVersion f2118484c5abc2fb6c4fde2749e54efad573b574
huggingface.co/axtr05/whisper-small-myanmar/tree/f2118484c5abc2fb6c4fde2749e54efad573b574

Verified Oct 31

Download files

Repository files

File inventory only; completeness is unverified.

Model artifacts

1 artifact

model.safetensors

safetensors · FP32 (unverified) · 922 MB · SHA-256 faa0ef26bfc1…0c84 · Hugging Face

Download file

Source excerpts

3 excerpts
ReadmeAug 15huggingface.co/axtr05/whisper-small-myanmar

--- license: apache-2.0 base_model: openai/whisper-small tags: - generated_from_trainer metrics: - wer model-index: - name: whisper-small-myanmar results: [] datasets: - chuuhtetnaing/myanmar-speech-dataset-openslr-80 language: - my pipeline_tag: automatic-speech-recognition library_name: transformers --- <!-- 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-myanmar This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the [chuuhtetnaing/myanmar-speech-dataset-openslr-80](https://huggingface.co/datasets/chuuhtetnaing/myanmar-speech-dataset-openslr-80) dataset. It achieves the following results on the evaluation set: - Loss: 0.1904 - Wer: 49.0650 ## Usage ```python from datasets import Audio, load_dataset from transformers import pipeline # Load a sample audio dataset = load_dataset("chuuhtetnaing/myanmar-speech-dataset-openslr-80") dataset = dataset.cast_column("audio", Audio(sampling_rate=16000)) test_dataset = dataset['test'] input_speech = test_dataset[42]['audio'] pipe = pipeline(model='chuuhtetnaing/whisper-small-myanmar') output = pipe(input_speech, generate_kwargs={"language": "myanmar", "task": "transcribe"}) print(output['text']) # ကျွန်မ ပြည်ပ မှာ ပညာသင် တော့ စာမေးပွဲ ကို တပတ်တခါ စစ်တယ် ``` ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 0.0003 - train_batch_size: 64 - eval_batch_size: 64 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 200 - num_epochs: 30 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 1.2566 | 1.0 | 36 | 0.8893 | 215.0045 | | 0.8862 | 2.0 | 72 | 0.6243 | 388.6465 | | 0.3546 | 3.0 | 108 | 0.2046 | 316.8744 | | 0.1839 | 4.0 | 144 | 0.1695 | 81.3001 | | 0.1198 | 5.0 | 180 | 0.1385 | 63.8914 | | 0.0969 | 6.0 | 216 | 0.1583 | 66.0285 | | 0.084 | 7.0 | 252 | 0.1539 | 70.6589 | | 0.0628 | 8.0 | 288 | 0.1603 | 61.3090 | | 0.0565 | 9.0 | 324 | 0.1424 | 60.3295 | | 0.0355 | 10.0 | 360 | 0.1457 | 58.1478 | | 0.0299 | 11.0 | 396 | 0.1547 | 57.7916 | | 0.0183 | 12....

NoteOct 31huggingface.co/axtr05/whisper-small-myanmar

Source context: 15 downloads · 0 likes · Pipeline automatic-speech-recognition · Library transformers · Repo axtr05/whisper-small-myanmar

NoteAug 15huggingface.co/axtr05/whisper-small-myanmar

Source context: 8 downloads · 0 likes · Pipeline automatic-speech-recognition · Library transformers · Repo axtr05/whisper-small-myanmar