Tested torchaudio's integration of flashlight-text and kenlm packages:
https://huggingface.co/spaces/Yehor/w2v-bert-uk-v2.1-lm-demo
#ai #speech
https://huggingface.co/spaces/Yehor/w2v-bert-uk-v2.1-lm-demo
#ai #speech
huggingface.co
Speech-to-Text for Ukrainian v2.1 (W2V-BERT 2.0) with LM - a Hugging Face Space by Yehor
Upload or record Ukrainian audio to get a text transcription. The app handles files up to 60 seconds long and provides transcriptions using a specialized model.
Made a simple Python script to generate Argilla project for audio annotation from a dataset:
https://github.com/egorsmkv/argilla-audio-annotation
https://github.com/egorsmkv/argilla-audio-annotation
GitHub
GitHub - egorsmkv/argilla-audio-annotation: Audio Annotation in Argilla - a python script creates an audio dataset to the Argilla…
Audio Annotation in Argilla - a python script creates an audio dataset to the Argilla instance - egorsmkv/argilla-audio-annotation
doing something
goreleaser supports Rust builds #go #rust
thw, goreleaser can build docker images - really useful to pipeline all steps together
doing something
https://github.com/cross-rs/cross/blob/main/README.md#supported-targets https://github.com/cross-rs/cross-toolchains https://github.com/rust-cross/cargo-xwin #rust
GitHub
GitHub - crs-org/ouch-releases: ouch releases for different platforms using GoReleaser
ouch releases for different platforms using GoReleaser - crs-org/ouch-releases
doing something
https://github.com/crs-org/ouch-releases Used cross to make this repo #rust
Applied new knowledge to my app that extracts audio from parquet/arrow files (standard way to pack audio data in Hugging Face datasets):
https://github.com/crs-org/extract-audio/releases/tag/v0.2.0
#rust
https://github.com/crs-org/extract-audio/releases/tag/v0.2.0
#rust