Seunghyun Lee

Support other embedding models

1

[static-retrieval-mrl-en-v1](https://huggingface.co/sentence-transformers/static-retrieval-mrl-en-v1) seems decent static embedding for on-device compute. Or maybe some other models like [M2V_base_glove](https://huggingface.co/minishlab/M2V_base_glove) would be good. Is there plan to add these models?

whisper: detect multiple languages

1

Log all language probs by descending order, when `-dl` option is used. Related issue: https://github.com/ggml-org/whisper.cpp/issues/3018 ### Example ``` % ./build/bin/whisper-cli -m ./models/ggml-tiny.bin -f ./samples/jfk.wav -dl ... whisper_full_with_state: auto-detected language: en...

Related issue: https://github.com/ggml-org/whisper.cpp/issues/1056 Add a `--file_mode` option to the stream example. This feature works together with the `--file` option. - `raw`: Writes every iteration, so it captures pre‑newline interim tokens,...