metadata
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
datasets:
- audiofolder
metrics:
- accuracy
model-index:
- name: Pak-Speech-Processing/urdu-emotion-whisper
results:
- task:
name: Audio Classification
type: audio-classification
dataset:
name: Pak-Speech-Processing/urdu-emotions
type: audiofolder
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.9166666666666666
Pak-Speech-Processing/urdu-emotion-whisper
This model is a fine-tuned version of openai/whisper-medium on the Pak-Speech-Processing/urdu-emotions dataset. It achieves the following results on the evaluation set:
- Loss: 0.5604
- Accuracy: 0.9167
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: 5e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.0018 | 1.0 | 120 | 2.0096 | 0.6667 |
| 0.5139 | 2.0 | 240 | 0.8303 | 0.8667 |
| 0.6903 | 3.0 | 360 | 0.8813 | 0.8833 |
| 0.0006 | 4.0 | 480 | 0.3012 | 0.95 |
| 1.5207 | 5.0 | 600 | 0.6310 | 0.8833 |
| 0.0005 | 6.0 | 720 | 0.5993 | 0.9 |
| 0.0004 | 7.0 | 840 | 0.3247 | 0.9167 |
| 0.0001 | 8.0 | 960 | 0.5303 | 0.9167 |
| 0.0001 | 9.0 | 1080 | 0.5530 | 0.9167 |
| 0.0001 | 10.0 | 1200 | 0.5604 | 0.9167 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2