Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cannot convert onnx to engine #77

Open
AIisCool opened this issue Dec 16, 2023 · 1 comment
Open

Cannot convert onnx to engine #77

AIisCool opened this issue Dec 16, 2023 · 1 comment

Comments

@AIisCool
Copy link

[12/15/2023-19:46:53] [V] [TRT] Registering layer: Add_121 for ONNX node: Add_121
[12/15/2023-19:46:53] [V] [TRT] Registering tensor: 186 for ONNX tensor: 186
[12/15/2023-19:46:53] [V] [TRT] Add_121 [Add] outputs: [186 -> (-1, 64, -1, -1)[FLOAT]], 
[12/15/2023-19:46:53] [V] [TRT] Static check for parsing node: Add_122 [Add]
[12/15/2023-19:46:53] [V] [TRT] Parsing node: Add_122 [Add]
[12/15/2023-19:46:53] [V] [TRT] Searching for input: 186
[12/15/2023-19:46:53] [V] [TRT] Searching for input: 65
[12/15/2023-19:46:53] [V] [TRT] Add_122 [Add] inputs: [186 -> (-1, 64, -1, -1)[FLOAT]], [65 -> (-1, 64, -1, -1)[FLOAT]], 
[12/15/2023-19:46:53] [V] [TRT] Registering layer: Add_122 for ONNX node: Add_122
[12/15/2023-19:46:53] [V] [TRT] Registering tensor: 187 for ONNX tensor: 187
[12/15/2023-19:46:53] [V] [TRT] Add_122 [Add] outputs: [187 -> (-1, 64, -1, -1)[FLOAT]], 
[12/15/2023-19:46:53] [V] [TRT] Static check for parsing node: Conv_123 [Conv]
[12/15/2023-19:46:53] [V] [TRT] Parsing node: Conv_123 [Conv]
[12/15/2023-19:46:53] [V] [TRT] Searching for input: 187
[12/15/2023-19:46:53] [V] [TRT] Searching for input: m_tail.weight
[12/15/2023-19:46:53] [V] [TRT] Conv_123 [Conv] inputs: [187 -> (-1, 64, -1, -1)[FLOAT]], [m_tail.weight -> (3, 64, 3, 3)[FLOAT]], 
[12/15/2023-19:46:53] [V] [TRT] Convolution input dimensions: (-1, 64, -1, -1)
[12/15/2023-19:46:53] [V] [TRT] Registering layer: Conv_123 for ONNX node: Conv_123
[12/15/2023-19:46:53] [V] [TRT] Using kernel: (3, 3), strides: (1, 1), prepadding: (1, 1), postpadding: (1, 1), dilations: (1, 1), numOutputs: 3
[12/15/2023-19:46:53] [V] [TRT] Convolution output dimensions: (-1, 3, -1, -1)
[12/15/2023-19:46:53] [V] [TRT] Registering tensor: output_0 for ONNX tensor: output
[12/15/2023-19:46:53] [V] [TRT] Conv_123 [Conv] outputs: [output -> (-1, 3, -1, -1)[FLOAT]], 
[12/15/2023-19:46:53] [V] [TRT] Marking output_0 as output: output
[12/15/2023-19:46:53] [I] Finished parsing network model. Parse time: 2.17735
[12/15/2023-19:46:53] [I] Set shape of input tensor input for optimization profile 0 to: MIN=1x3x574x706 OPT=1x3x574x706 MAX=1x3x574x706
[12/15/2023-19:46:53] [V] [TRT] Trying to set exclusive file lock C:/VapourSynth/vsmlrt/models/dpir\f82cfa0a.engine.cache.lock

[12/15/2023-19:46:53] [W] [TRT] Could not read timing cache from: C:/VapourSynth/vsmlrt/models/dpir\f82cfa0a.engine.cache. A new timing cache will be generated and written.
[12/15/2023-19:46:53] [E] Error[4]: [network.cpp::nvinfer1::Network::validate::3640] Error Code 4: Internal Error (input: for dimension number 1 in profile 0 does not match network definition (got min=3, opt=3, max=3), expected min=opt=max=4).)
[12/15/2023-19:46:53] [E] Engine could not be created from network
[12/15/2023-19:46:53] [E] Building engine failed
[12/15/2023-19:46:53] [E] Failed to create engine from model or file.
[12/15/2023-19:46:53] [E] Engine set up failed
&&&& FAILED TensorRT.trtexec [TensorRT v9100] # C:/VapourSynth/vsmlrt\vsmlrt-cuda\trtexec --onnx=C:/VapourSynth/vsmlrt/models/dpir/drunet_deblocking_color.onnx
@WolframRhodium
Copy link
Contributor

WolframRhodium commented Dec 16, 2023

DPIR is different in that it requires an additional input to specify the noise strength, so the number of input channel is 4 instead of 3.

In addition, it also requires dimensions to be multiples of 8, so it should be trtexec --shapes=1x4x576x712.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants