mr-large-v2 (CTranslate2)

A CTranslate2 conversion of DrishtiSharma/whisper-large-v2-marathi for faster-whisper, as downloaded by LiveWhisper, a voice dictation app for Windows.

All credit for the model goes to its original authors. This repository only changes the file format (float16 weights, plus tokenizer.json and preprocessor_config.json that faster-whisper needs). The licence is the original's: apache-2.0.

Large-v2 sized: as fast as large-v3 on the same clips (2.5x realtime), allow ~6 GB of VRAM with the main model. Converting peaks near 11.5 GB of memory.

Measured in LiveWhisper

metric result
FLEURS word error 47.2% (large-v3 78.7%)
delivered Minglish 51.6% (large-v3 73.2%)
from faster_whisper import WhisperModel
model = WhisperModel("itsskofficial/livewhisper-mr-large-v2", device="cuda", compute_type="int8_float16")
Downloads last month
14
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for itsskofficial/livewhisper-mr-large-v2

Finetuned
(1)
this model