the TRTIS can not load the trt model about server HOT 4 CLOSED

The UFF file itself is not a TRT plan. You need to import the UFF file into TRT and from that generate a plan. You can then use that plan file in your model store. I think this sample shows what you are trying to do: https://docs.nvidia.com/deeplearning/sdk/tensorrt-developer-guide/index.html#mnist_uff_sample

from server.

thanks! it works by serializing the engine

from server.

The UFF file itself is not a TRT plan. You need to import the UFF file into TRT and from that generate a plan. You can then use that plan file in your model store. I think this sample shows what you are trying to do: https://docs.nvidia.com/deeplearning/sdk/tensorrt-developer-guide/index.html#mnist_uff_sample

We got the error when transforming UFF File to TensorRT Engine (TensorRT 5.0.2.6):
"[TensorRT] ERROR: UFFParser: Parser error: add: The input to the Scale Layer is required to have a minimum of 3 dimensions."
our model details as:
input nodes ==:
[name: "Placeholder"
op: "Placeholder"
attr {
key: "_output_shapes"
value {
list {
shape {
dim {
size: -1
}
dim {
size: 1000
}
}
}
}
}

output nodes ==:
[name: "add"
op: "Add"
input: "MatMul_2"
input: "b"
attr {
key: "T"
value {
type: DT_FLOAT
}
}
attr {
key: "_output_shapes"
value {
list {
shape {
dim {
size: -1
}
dim {
size: 2
}
}
}
}
}
]

from server.

UFF expects tensors in the network to be 3D. There will be some documentation on this limitation in the (future) 5.1 release of TensorRT.

from server.