LLauraGPTHarden SenseVoice streaming server

558bd67c创建于 8月13日历史提交

文件	最后提交记录	最后更新时间
golden	test: numerical regression harness (frozen golden vs ggml/VAD/CIF/CTC output) (#301) * test: numerical regression harness (frozen golden vs ggml/VAD/CIF/CTC output) Adds tests/ — runs each runtime tool on a fixed 6 s clip and diffs against frozen golden output, catching regressions in the ggml graphs, the FSMN-VAD state machine, the CIF predictor and CTC decode. - tests/run_regression.sh: auto-detects which tools are built; VAD model is auto-fetched (1.7 MB), ASR GGUFs tested when present or with RUN_FULL=1 (downloads from HF). Non-zero exit on any mismatch. BIN_DIR/MODELS_DIR overridable. - tests/sample.wav (~6 s) + tests/golden/.txt: golden captured on Linux x86-64 with the f16 GGUFs from FunAudioLLM/-GGUF. Verified locally: all present tools PASS; default mode fetches VAD + skips absent models. * test: golden as text (binaries now detok in-process; merge after in-binary-detok) * test: fix sensevoice golden (correct transcription text)	3 个月前
README.md	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前
run_regression.sh	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前
run_server_contract.sh	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前
run_server_smoke.sh	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前
sample.wav	test: numerical regression harness (frozen golden vs ggml/VAD/CIF/CTC output) (#301) * test: numerical regression harness (frozen golden vs ggml/VAD/CIF/CTC output) Adds tests/ — runs each runtime tool on a fixed 6 s clip and diffs against frozen golden output, catching regressions in the ggml graphs, the FSMN-VAD state machine, the CIF predictor and CTC decode. - tests/run_regression.sh: auto-detects which tools are built; VAD model is auto-fetched (1.7 MB), ASR GGUFs tested when present or with RUN_FULL=1 (downloads from HF). Non-zero exit on any mismatch. BIN_DIR/MODELS_DIR overridable. - tests/sample.wav (~6 s) + tests/golden/.txt: golden captured on Linux x86-64 with the f16 GGUFs from FunAudioLLM/-GGUF. Verified locally: all present tools PASS; default mode fetches VAD + skips absent models. * test: golden as text (binaries now detok in-process; merge after in-binary-detok) * test: fix sensevoice golden (correct transcription text)	3 个月前
server_utils_test.cpp	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前
stream_client.py	Harden SenseVoice streaming server Signed-off-by: LauraGPT <lauragpt@users.noreply.github.com>	1 个月前

Regression tests

Run each runtime tool on a fixed clip (sample.wav, ~6 s) and diff the output against the frozen golden in golden/. Catches regressions in the ggml graphs, the FSMN-VAD state machine, the CIF predictor and CTC decode.

Run

# from runtime/llama.cpp/, after building (cmake --build build):
./tests/run_regression.sh              # VAD (tiny model auto-fetched) + any tool whose GGUF is already local
RUN_FULL=1 ./tests/run_regression.sh   # also download the ASR GGUFs from Hugging Face and test every tool

# model-free server CLI, naming, and resource-limit checks:
BIN_DIR="$PWD/build/bin" ./tests/run_server_contract.sh

# REST formats, SSE, malformed/limited WebSockets, and realtime transcription:
BIN_DIR="$PWD/build/bin" MODEL_GGUF=/path/to/sensevoice.gguf \
  VAD_GGUF=/path/to/fsmn-vad.gguf ./tests/run_server_smoke.sh

BIN_DIR / MODELS_DIR override where binaries and GGUFs are found. Exit code is non-zero if any test fails; tools with no binary or no model are skipped.

Golden

Captured on Linux x86-64 (the reference platform) with the f16 GGUFs published at FunAudioLLM/*-GGUF. Update a golden file only with a deliberate, reviewed output change.