Wav2vec2 Xlsr 53 Russian Aligner Coreml
This repository contains a fixed-contract Core ML conversion intended for offline Russian speech alignment.
Source du modèle
- Type
- Other
- Taille du fichier
- 1,18 GB
- Version
- vRevision 2cab913
- Mis à jour
- 27 sept. 2026
- Téléchargements
- 0
- Source
- Hugging Face
Description de la source
Russian Wav2Vec2 alignment model for Core ML
This repository contains a fixed-contract Core ML conversion intended for offline Russian speech alignment.
Provenance
- Upstream model: jonatasgrosman/wav2vec2-large-xlsr-53-russian
- Immutable upstream revision:
2329100508896c6d9b157019803ab5601e6f3406 - Upstream license:
apache-2.0
Conversion does not change the upstream license. Review its terms before use or redistribution.
Core ML contract
- Input
waveform: shape[1,960000], 16 kHz mono waveform - Output
emissions: shape[1,2999,39], log probabilities - Sample rate: 16000 Hz
The accompanying conversion manifest records tool versions and SHA-256 hashes. The validation report must pass before this repository can be published.
Sources
1 source
Sources
1 sourceVérifié 30 août
Artefacts du modèle
1 artefact
Artefacts du modèle
1 artefactwav2vec2_xlsr53_russian_aligner.mlpackage/Data/com.apple.CoreML/weights/weight.bin
bin · 1,18 GB · SHA-256 977175f5d942…690c · Hugging Face
Download fileExtraits de sources
2 extraits
Extraits de sources
2 extraits--- license: apache-2.0 tags: - coreml - wav2vec2 - audio - automatic-speech-recognition --- # Russian Wav2Vec2 alignment model for Core ML This repository contains a fixed-contract Core ML conversion intended for offline Russian speech alignment. ## Provenance - Upstream model: [jonatasgrosman/wav2vec2-large-xlsr-53-russian](https://huggingface.co/jonatasgrosman/wav2vec2-large-xlsr-53-russian) - Immutable upstream revision: `2329100508896c6d9b157019803ab5601e6f3406` - Upstream license: `apache-2.0` Conversion does not change the upstream license. Review its terms before use or redistribution. ## Core ML contract - Input `waveform`: shape `[1,960000]`, 16 kHz mono waveform - Output `emissions`: shape `[1,2999,39]`, log probabilities - Sample rate: 16000 Hz The accompanying conversion manifest records tool versions and SHA-256 hashes. The validation report must pass before this repository can be published.
Source context: 0 downloads · 0 likes · Pipeline automatic-speech-recognition · Repo vadimsuhanov/wav2vec2-xlsr-53-russian-aligner-coreml
