seamless_communication/ggml

unity.cpp

Introduction

GGML is an open source library in C to enable large model inference on various hardware platforms. We implemented unity.cpp in ggml. Now it supports SeamlessM4T model for X2T tasks - Speech-to-text translation (S2TT), Acoustic speech recognition (ASR), Text-to-text translation (T2TT).

The project is still active in development. Contributions are welcome!

Build

To build the interactive console for S2TT & ASR,

cd seamless_communication/ggml
mkdir build; cd build
cmake -DGGML_OPENBLAS=ON \
    -DBUILD_SHARED_LIBS=On \
	  -DCMAKE_BUILD_TYPE=Release \
	  -DCMAKE_CXX_FLAGS="-g2 -fno-omit-frame-pointer" \
    ..
make -j4 unity # Interactive Console

For more build commands see Makefile.

CLI usage

Command to launch an interactive console for S2TT & ASR, note that the model already includes vocabulary needed to detokenize.

OPENBLAS_NUM_THREADS=8 ./bin/unity --model seamlessM4T_medium.ggml

In the console, enter the path of local waveform file and target language, separated by space. Note that the first run would include some “warm up” time so could be slow.

Converted ggml models could be downloaded from |SeamlessM4T_large | SeamlessM4T_medium | NLLB_dense_1b | NLLB_distill_600m | |-------- | -------- | ------- | ------- | | model | model | model | model

For more details of NLLB models, please check https://github.com/facebookresearch/fairseq/tree/nllb.

Fairseq2 model conversion

Models from fairseq2 checkpoints could be converted to ggml automatically with ggml_convert.py.

python ggml_convert.py -m MODEL_NAME

where MODEL_NAME corresponds to asset cards in fairseq2 / seamless_communication, e.g. seamlessM4T_medium, seamlessM4T_large

Python bindings

We also utilize ggml python bindings for better dev experience. For examples of running unity.cpp in python, refer to tests in test_unity_cpp.py.

[Optional]Dependencies

OpenBLAS

We strongly suggest building with OpenBLAS, as we've seen 8x speedup on test machine.

libsndfile

This is needed only for the console to load waveform, but not the library.