Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

fix: accept float32 dtype in realtime websocket server (#3204)

zhifu gao505455eb6477
diff --git a/examples/industrial_data_pretraining/fun_asr_nano/serve_realtime_ws.py b/examples/industrial_data_pretraining/fun_asr_nano/serve_realtime_ws.py
index 43e785329..f6e40cce5 100644
--- a/examples/industrial_data_pretraining/fun_asr_nano/serve_realtime_ws.py
+++ b/examples/industrial_data_pretraining/fun_asr_nano/serve_realtime_ws.py
@@ -585,7 +585,12 @@ def build_arg_parser():
parser.add_argument("--spk-model", type=str, default="iic/speech_eres2netv2_sv_zh-cn_16k-common")
parser.add_argument("--hotword-file", type=str, default="热词列表")
parser.add_argument("--language", type=str, default=None, help="Language hint (e.g. 中文, English, 日本語)")
- parser.add_argument("--dtype", type=str, default="bf16", choices=["bf16", "fp16", "fp32"])
+ parser.add_argument(
+ "--dtype",
+ type=lambda value: "fp32" if value == "float32" else value,
+ default="bf16",
+ choices=["bf16", "fp16", "fp32", "float32"],
+ )
parser.add_argument("--tensor-parallel-size", type=int, default=1)
parser.add_argument("--gpu-memory-utilization", type=float, default=0.8)
parser.add_argument("--max-model-len", type=int, default=2048)
diff --git a/tests/test_realtime_ws_service.py b/tests/test_realtime_ws_service.py
index 1ed0231a4..62396a913 100644
--- a/tests/test_realtime_ws_service.py
+++ b/tests/test_realtime_ws_service.py
@@ -36,6 +36,14 @@ def test_cli_defaults_disable_speaker_and_bound_partial_window():
assert args.partial_window_sec == 15.0
+def test_cli_accepts_float32_dtype_alias():
+ module = load_service_module()
+
+ args = module.build_arg_parser().parse_args(["--dtype", "float32"])
+
+ assert args.dtype == "fp32"
+
+
def test_websocket_keepalive_kwargs_are_configurable():
module = load_service_module()