Whisper Large V3 Russian Podlodka
| Property | Value |
|---|---|
| Parameter Count | 1.54B |
| License | Apache 2.0 |
| Tensor Type | F32 |
| Author | bond005 |
What is whisper-large-v3-ru-podlodka?
This is a specialized Russian speech recognition model based on OpenAI's Whisper Large V3 architecture, fine-tuned specifically for Russian language processing. The model has been optimized using multiple Russian datasets including Taiga Speech V2, Podlodka Speech, and Russian LibriSpeech.
Implementation Details
The model implements a transformer-based architecture with 1.54 billion parameters, utilizing PyTorch and Safetensors for efficient processing. It achieves impressive Word Error Rates (WER) of 10.987% without punctuation on the Podlodka.io dataset and 9.795% on Russian LibriSpeech.
- Trained on multiple high-quality Russian speech datasets
- Optimized for both clean and noisy audio conditions
- Supports punctuation and capitalization in transcriptions
- Implements state-of-the-art Whisper V3 architecture
Core Capabilities
- High-accuracy Russian speech recognition
- Robust performance in various acoustic conditions
- Support for punctuation and case-sensitive transcription
- Achieves ~21% WER with punctuation and ~11% WER without punctuation
Frequently Asked Questions
Q: What makes this model unique?
This model stands out due to its specialized optimization for Russian language processing and its comprehensive training on multiple Russian speech datasets, making it particularly effective for Russian speech recognition tasks.
Q: What are the recommended use cases?
The model is ideal for Russian speech transcription tasks, particularly in scenarios requiring high accuracy such as podcast transcription, meeting recordings, and general speech-to-text applications where Russian language support is crucial.
