--- language: - fi license: apache-2.0 tags: - whisper-event - finnish - speech-recognition datasets: - mozilla-foundation/common_voice_11_0 - google/fleurs metrics: - wer - cer model-index: - name: Whisper Large V3 Finnish results: - task: name: Automatic Speech Recognition type: automatic-speech-recognition dataset: name: Common Voice 11.0 type: mozilla-foundation/common_voice_11_0 config: fi split: test args: fi metrics: - name: Wer type: wer value: 8.23 - name: Cer type: cer value: 1.43 - task: name: Automatic Speech Recognition type: automatic-speech-recognition dataset: name: FLEURS type: google/fleurs config: fi_fi split: test args: fi_fi metrics: - name: Wer type: wer value: 8.21 - name: Cer type: cer value: 3.23 library_name: transformers pipeline_tag: automatic-speech-recognition --- This is a conversion of [Finnish-NLP/whisper-large-finnish-v3](https://huggingface.co/Finnish-NLP/whisper-large-finnish-v3) into faster-whisper format.
We improve from our previously finetuned Whisper V2 model in the following mannerhttps://huggingface.co/Finnish-NLP/whisper-large-v2-finnish
CV11 (Common Voice 11 test set) WER (Word error rate) 10.42 --> 8.23
Fleurs (A speech recognition test set by Google) WER (Word error rate) 10.20 --> 8.21
Model was trained on Nvidia RTX4080 for 32k steps with batch size 8, gradient accumulation 2