YoloV8 export still does not work

[mgabell]

The export seems to work. Training goes fine. The prediction with ultralytics works fine. But deepstream v8 does not create boxes. There must be an issue still with the export. I use the export_yolov8.py and run as:

python3 export_yolov8.py -w myYoloModel.pt --dynamic.

This will generate a onnx model (very fast). But using it does not work. No erros, the video plays, but does not produce any defects. I sometimes get a model working. So what will it depend on? epochs?

WARNING: Deserialize engine failed because file path: /home/aiadmin/Development/deepstream-yolov8-evaluation/model_b1_gpu0_fp16.engine open error
0:00:03.580736845  7375     0x31891b20 WARN                 nvinfer gstnvinfer.cpp:677:gst_nvinfer_logger:<primary-inference> NvDsInferContext[UID 1]: Warning from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:1897> [UID = 1]: deserialize engine from file :/home/aiadmin/Development/deepstream-yolov8-evaluation/model_b1_gpu0_fp16.engine failed
0:00:03.754231754  7375     0x31891b20 WARN                 nvinfer gstnvinfer.cpp:677:gst_nvinfer_logger:<primary-inference> NvDsInferContext[UID 1]: Warning from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2002> [UID = 1]: deserialize backend context from engine from file :/home/aiadmin/Development/deepstream-yolov8-evaluation/model_b1_gpu0_fp16.engine failed, try rebuild
0:00:03.754333067  7375     0x31891b20 INFO                 nvinfer gstnvinfer.cpp:680:gst_nvinfer_logger:<primary-inference> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1923> [UID = 1]: Trying to create engine from model files
WARNING: [TRT]: onnx2trt_utils.cpp:375: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
WARNING: [TRT]: Tensor DataType is determined at build time for tensors not marked as input or output.

Building the TensorRT Engine

Building complete

0:04:08.014390407  7375     0x31891b20 INFO                 nvinfer gstnvinfer.cpp:680:gst_nvinfer_logger:<primary-inference> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1955> [UID = 1]: serialize cuda engine to file: /home/aiadmin/Development/deepstream-yolov8-evaluation/model_b1_gpu0_fp32.engine successfully
WARNING: [TRT]: The getMaxBatchSize() function should not be used with an engine built from a network created with NetworkDefinitionCreationFlag::kEXPLICIT_BATCH flag. This function will always return 1.
INFO: [Implicit Engine Info]: layers num: 4
0   INPUT  kFLOAT input           3x640x640       
1   OUTPUT kFLOAT boxes           8400x4          
2   OUTPUT kFLOAT scores          8400x1          
3   OUTPUT kFLOAT classes         8400x1          

0:04:08.261001291  7375     0x31891b20 INFO                 nvinfer gstnvinfer_impl.cpp:328:notifyLoadModelStatus:<primary-inference> [UID 1]: Load new model:config_infer_primary_yoloV8.txt sucessfully
Decodebin child added: source 

Decodebin child added: decodebin0 

**PERF:  {'stream0': 0.0} 
Decodebin child added: qtdemux0 
Decodebin child added: multiqueue0 
Decodebin child added: h264parse0 
Decodebin child added: capsfilter0 
Decodebin child added: nvv4l2decoder0 

Opening in BLOCKING MODE 
NvMMLiteOpen : Block : BlockType = 261 
NVMEDIA: Reading vendor.tegra.display-size : status: 6 
NvMMLiteBlockCreate : Block : BlockType = 261 
In cb_newpad

gstname= video/x-raw
features= <Gst.CapsFeatures object at 0xffff6b680ca0 (GstCapsFeatures at 0x8a1cc6e0)>

[mgabell]

It works now...

[xxiMiaxx]

It works now...

okay but what was the problem and how did you solve it.

[mgabell]

Sorry to be honest I dont know. I changed to yolov5. Installed required packages. I then created a trained model with ultralytics. I also made sure to follow marcos instruction to the letter. I used HIS export tool and then remembered to move the labels.txt and onnx model to the correct folder for deepstream. Then I deleted the engine file and ensured that the name in the config file and the engine file was correct.

See if that works. I have work my "#¤%"# off to get this work I will do what I can to assist.