onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from moss-moon-003-sft-plugin failed:Protobuf parsing failed

参考 https://github.com/OpenLMLab/MOSS_Vortex 换了个模型

2023-05-11 11:04:36,394 - 82 - INFO - mosec_server.py:357 - [MOSEC] [INIT] Initializing model on device=3
2023-05-11 11:04:36,423 - 124 - INFO - mosec_server.py:357 - [MOSEC] [INIT] Initializing model on device=4
2023-05-11 11:04:36,443 - 79 - INFO - mosec_server.py:357 - [MOSEC] [INIT] Initializing model on device=0
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:36,621 - 81 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cuda
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
{"timestamp": "2023-05-11T11:04:36.715112+0800", "level": "INFO", "fields": {"message": "<1|Preprocess|6> socket connected to /tmp/mosec_a635e0d/ipc_1.socket"}, "target": "protocol.py:open"}
2023-05-11 11:04:36,715 - 78 - INFO - protocol.py:141 - <1|Preprocess|6> socket connected to /tmp/mosec_a635e0d/ipc_1.socket
{"timestamp": "2023-05-11T11:04:36.742202+0800", "level": "INFO", "fields": {"message": "<1|Preprocess|1> socket connected to /tmp/mosec_a635e0d/ipc_1.socket"}, "target": "protocol.py:open"}
2023-05-11 11:04:36,742 - 53 - INFO - protocol.py:141 - <1|Preprocess|1> socket connected to /tmp/mosec_a635e0d/ipc_1.socket
{"timestamp": "2023-05-11T11:04:36.760997+0800", "level": "INFO", "fields": {"message": "<1|Preprocess|5> socket connected to /tmp/mosec_a635e0d/ipc_1.socket"}, "target": "protocol.py:open"}
2023-05-11 11:04:36,760 - 67 - INFO - protocol.py:141 - <1|Preprocess|5> socket connected to /tmp/mosec_a635e0d/ipc_1.socket
{"timestamp": "2023-05-11T11:04:36.798910+0800", "level": "INFO", "fields": {"message": "<1|Preprocess|4> socket connected to /tmp/mosec_a635e0d/ipc_1.socket"}, "target": "protocol.py:open"}
2023-05-11 11:04:36,798 - 64 - INFO - protocol.py:141 - <1|Preprocess|4> socket connected to /tmp/mosec_a635e0d/ipc_1.socket
{"timestamp": "2023-05-11T11:04:36.883819+0800", "level": "INFO", "fields": {"message": "<1|Preprocess|3> socket connected to /tmp/mosec_a635e0d/ipc_1.socket"}, "target": "protocol.py:open"}
2023-05-11 11:04:36,883 - 61 - INFO - protocol.py:141 - <1|Preprocess|3> socket connected to /tmp/mosec_a635e0d/ipc_1.socket
2023-05-11 11:04:37,023 - 81 - INFO - mosec_server.py:366 - [MOSEC] [INIT] ONNX Model Loading
Process SpawnProcess-9:
Traceback (most recent call last):
  File "/usr/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
    self.run()
  File "/usr/lib/python3.8/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "/usr/local/lib/python3.8/dist-packages/mosec/coordinator.py", line 81, in __init__
    self.worker = worker()
  File "/mosec/mosec_server.py", line 367, in __init__
    self.ort_session = ort.InferenceSession(self.model_path, ort.SessionOptions(),providers=['CUDAExecutionProvider', 'CPUExecutionProvider'])
  File "/usr/local/lib/python3.8/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 360, in __init__
    self._create_inference_session(providers, provider_options, disabled_optimizers)
  File "/usr/local/lib/python3.8/dist-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 397, in _create_inference_session
    sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from moss-moon-003-sft-plugin failed:Protobuf parsing failed.
2023-05-11 11:04:38,155 - 126 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cpu
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:38,162 - 80 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cuda
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:38,170 - 124 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cpu
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:38,216 - 82 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cuda
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:38,224 - 79 - INFO - mosec_server.py:359 - [MOSEC] [INIT] Using computing device: cuda
Explicitly passing a `revision` is encouraged when loading a model with custom code to ensure no malicious code has been contributed in a newer revision.
2023-05-11 11:04:38,602 - 126 - INFO - mosec_server.py:366 - [MOSEC] [INIT] ONNX Model Loading
2023-05-11 11:04:38,606 - 80 - INFO - mosec_server.py:366 - [MOSEC] [INIT] ONNX Model Loading
2023-05-11 11:04:38,698 - 79 - INFO - mosec_server.py:366 - [MOSEC] [INIT] ONNX Model Loading
2023-05-11 11:04:38,712 - 124 - INFO - mosec_server.py:366 - [MOSEC] [INIT] ONNX Model Loading

Start at mosec_server.py line 367, where ort.InferenceSession loads the configured model path, and compare the substituted model with the expected ONNX input. Verify the model can be parsed before initialization completes; done means the model loads without the INVALID_PROTOBUF error.