[Bug] [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from yolov8n.enginefailed:Protobuf parsing failed.

[Frank-fjh]

Describe the issue

I want to use onnxruntime with tensorrt ep but failed, this is my code

from ultralytics import YOLO
model = YOLO("yolov8n.pt")
model.export(format="engine") 
sess = ort.InferenceSession('./yolov8n.engine', providers=['TensorrtExecutionProvider'])

I can't load the model but i can use ultralytics.YOLO.predict to run the model with tensorrt.

---------------------------------------------------------------------------
InvalidProtobuf                           Traceback (most recent call last)
Cell In[3], [line 1](vscode-notebook-cell:?execution_count=3&line=1)
----> [1](vscode-notebook-cell:?execution_count=3&line=1) sess = ort.InferenceSession('./yolov8n.engine', 
      [2](vscode-notebook-cell:?execution_count=3&line=2)                             providers=['TensorrtExecutionProvider'])

File c:\Users\TL\anaconda3\envs\3_11\Lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py:419, in InferenceSession.__init__(self, path_or_bytes, sess_options, providers, provider_options, **kwargs)
    [416](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:416) disabled_optimizers = kwargs.get("disabled_optimizers")
    [418](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:418) try:
--> [419](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:419)     self._create_inference_session(providers, provider_options, disabled_optimizers)
    [420](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:420) except (ValueError, RuntimeError) as e:
    [421](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:421)     if self._enable_fallback:

File c:\Users\TL\anaconda3\envs\3_11\Lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py:480, in InferenceSession._create_inference_session(self, providers, provider_options, disabled_optimizers)
    [477](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:477) self._register_ep_custom_ops(session_options, providers, provider_options, available_providers)
    [479](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:479) if self._model_path:
--> [480](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:480)     sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
    [481](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:481) else:
    [482](file:///C:/Users/TL/anaconda3/envs/3_11/Lib/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py:482)     sess = C.InferenceSession(session_options, self._model_bytes, False, self._read_config_from_model)

InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from ./model.trt failed:Protobuf parsing failed.

Can anyone tell me what's wrong with it? Thanks.

To reproduce

Python-3.11.9
torch-2.4.0
CUDA:0 (NVIDIA GeForce RTX 4070 Laptop GPU, 8188MiB)
tenssort-gpu-1.19.0

Urgency

No response

Platform

Windows

OS Version

11

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.19.0

ONNX Runtime API

Python

Architecture

X64

Execution Provider

TensorRT

Execution Provider Library Version

CUDA11.4 TensorRT-10.0.1.6 cudnn-8.9.7

Model File

No response

Is this a quantized model?

No

[tianleiwu]

sess = ort.InferenceSession('./yolov8n.engine', providers=['TensorrtExecutionProvider'])

The input file for inference session shall be onnx model, not a TRT engine.

[Frank-fjh]

sess = ort.InferenceSession('./yolov8n.engine', providers=['TensorrtExecutionProvider'])

The input file for inference session shall be onnx model, not a TRT engine.

Thanks a lot!