W

Whisper Medicalv1

Developed by Crystalcareai
Distil-Whisper is a knowledge-distilled version of Whisper large-v3, focusing on English speech recognition, offering faster inference speeds while maintaining accuracy close to the original model.
Downloads 348
Release Time : 4/28/2024

Model Overview

This is a distilled model optimized for English speech recognition, based on OpenAI's Whisper large-v3, trained using large-scale pseudo-label knowledge distillation techniques, specifically optimized for accuracy in long-form audio transcription.

Model Features

Efficient Inference
6.3x faster than the original Whisper large-v3 and 1.1x faster than the previous distil-large-v2
Optimized Long-Form Transcription
Specially tuned to provide better long-form transcription accuracy when using sequential long-form algorithms
Compatibility with Mainstream Libraries
Designed to be compatible with popular libraries like Whisper cpp, Faster-Whisper, and OpenAI Whisper
Knowledge Distillation
Trained using large-scale pseudo-label knowledge distillation techniques, maintaining performance close to the original model

Model Capabilities

English Speech Recognition
Short-Form Audio Transcription (<30 seconds)
Long-Form Audio Transcription (>30 seconds)
Supports both sequential and chunked long-form transcription algorithms
Provides segment-level timestamps

Use Cases

Speech Transcription
Meeting Minutes Transcription
Convert meeting recordings into text transcripts
High accuracy, especially suitable for long-form meeting recordings
Podcast Transcription
Convert podcast audio content into text
Fast processing of long audio while maintaining high accuracy
Media Processing
Video Subtitle Generation
Automatically generate subtitles for video content
Efficient processing of long video content
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase