Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.
Python asraudiochineseemotion-recognitionfunasrmcp-servermultilingual-asropenai-compatible-apiparaformerpunctuationpytorchreal-time-asrspeaker-diarizationspeech-recognitionspeech-to-textstreaming-asrtranscriptionvllmvoice-activity-detectionwhisper-alternative
feat: support enforce-eager for realtime server (#3631)
Junhui Lida5a11b77eb6
diff --git a/funasr/bin/realtime_ws.py b/funasr/bin/realtime_ws.pyindex 5659b7cc1..f49d01dc4 100644--- a/funasr/bin/realtime_ws.py+++ b/funasr/bin/realtime_ws.py@@ -1333,6 +1333,7 @@ def load_models(args):tensor_parallel_size=getattr(args, 'tensor_parallel_size', 1),gpu_memory_utilization=getattr(args, 'gpu_memory_utilization', 0.8),max_model_len=getattr(args, 'max_model_len', 2048),+ enforce_eager=getattr(args, 'enforce_eager', False),)_vllm_engine = RealtimeBatchingEngine(engine,@@ -1679,6 +1680,11 @@ def build_arg_parser():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)+ parser.add_argument(+ "--enforce-eager",+ action="store_true",+ help="Disable vLLM compilation and CUDA graphs.",+ )parser.add_argument("--ws-ping-interval", type=float, default=20.0,help="WebSocket ping interval in seconds; <=0 disables keepalive pings.")parser.add_argument(