You can deploy this model and start earning money today!

KasuleTrevor/
cdli-whisper-en-ug-sunbird-lora-a40

Name: cdli-whisper-en-ug-sunbird-lora-a40
Availability: InStock

cdli/ugandan_english_nonstandard_speech_v1.0

Sunbird/asr-whisper-large-v3-salt

mit

Views

No views yet

cdli-whisper-en-ug-sunbird-lora-a40

This model is a fine-tuned version of Sunbird/asr-whisper-large-v3-salt on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.9183
Wer: 0.3700
Cer: 0.2687
Overall WER (normalized): 0.518
Overall CER (normalized): 0.364
Avg WER (normalized): 0.293
Avg CER (normalized): 0.207

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: 2e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 150
training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.7447	1.5445	250	0.9719	0.4008	0.2972
0.6838	3.0866	500	0.9344	0.3856	0.2837
0.6339	4.6311	750	0.9196	0.3737	0.2732
0.5693	6.1732	1000	0.9209	0.3822	0.2814
0.5792	7.7177	1250	0.9179	0.3709	0.2705
0.5807	9.2599	1500	0.9183	0.3700	0.2687

Framework versions

PEFT 0.19.1
Transformers 4.52.0
Pytorch 2.7.1+cu118
Datasets 3.6.0
Tokenizers 0.21.4

KasuleTrevor/cdli-whisper-en-ug-sunbird-lora-a40

cdli-whisper-en-ug-sunbird-lora-a40

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

KasuleTrevor/
cdli-whisper-en-ug-sunbird-lora-a40