You can deploy this model and start earning money today!

amelov/
whisper_large_v3_turbo_finetuned_for_mk

Name: whisper_large_v3_turbo_finetuned_for_mk
Availability: InStock

transformers

safetensors

whisper

automatic-speech-recognition

text-generation-inference

unsloth

fsicoli/common_voice_22_0

unsloth/whisper-large-v3-turbo

finetune

apache-2.0

endpoints_compatible

Views

No views yet

This is a finetuned whisper-large-v3-turbo model for Macedonian language. The dataset used for training and validation is fsicoli/common_voice_22_0. The finetuned model was tested by calculating WER and CER.

Metric	Base Model	Fine-tuned	Imporvement
WER	34.85	19.7	15.15
CER	11.4	4.76	6.64

Training Details

Base Model: openai/whisper-large-v3-turbo
Training Dataset: fsicoli/common_voice_22_0
Train Loss: 0.2941
Training Time: 196.65 minutes on T4 Google Colab

Uploaded finetuned model

Developed by: amelov
License: apache-2.0
Finetuned from model : unsloth/whisper-large-v3-turbo

This whisper model was trained 2x faster with Unsloth and Huggingface's TRL library.

amelov/whisper_large_v3_turbo_finetuned_for_mk

Training Details

Uploaded finetuned model

amelov/
whisper_large_v3_turbo_finetuned_for_mk