Method 'forward' is not defined error !

[MHmi1]

Hello, good time
i have exported jit script file using this code : https://github.com/k2-fsa/icefall/blob/master/egs/librispeech/ASR/zipformer/export.py
this is for my ASR project ,but i have this problem when i want to start Triton Server :

root@workstation-003:/workspace/Triton-ASR-Client/jit_config_files/model# tritonserver --model-repository=/workspace/Triton-ASR-Client/jit_config_files --http-port=8008 --grpc-port=8009 --metrics-port 8010
I0126 14:16:30.950672 2312 pinned_memory_manager.cc:277] "Pinned memory pool is created at '0x7f017e000000' with size 268435456"
I0126 14:16:30.951132 2312 cuda_memory_manager.cc:107] "CUDA memory pool is created on device 0 with size 67108864"
I0126 14:16:30.957236 2312 model_lifecycle.cc:472] "loading: model:1"
I0126 14:16:32.251723 2312 libtorch.cc:2547] "TRITONBACKEND_Initialize: pytorch"
I0126 14:16:32.251784 2312 libtorch.cc:2557] "Triton TRITONBACKEND API version: 1.19"
I0126 14:16:32.251805 2312 libtorch.cc:2563] "'pytorch' TRITONBACKEND API version: 1.19"
I0126 14:16:32.252781 2312 libtorch.cc:2596] "TRITONBACKEND_ModelInitialize: model (version 1)"
W0126 14:16:32.253609 2312 libtorch.cc:329] "skipping model configuration auto-complete for 'model': not supported for pytorch backend"
I0126 14:16:32.254088 2312 libtorch.cc:358] "Optimized execution is enabled for model instance 'model'"
I0126 14:16:32.254099 2312 libtorch.cc:377] "Cache Cleaning is disabled for model instance 'model'"
I0126 14:16:32.254111 2312 libtorch.cc:394] "Inference Mode is enabled for model instance 'model'"
I0126 14:16:32.254123 2312 libtorch.cc:413] "cuDNN is enabled for model instance 'model'"
I0126 14:16:32.254233 2312 libtorch.cc:2640] "TRITONBACKEND_ModelInstanceInitialize: model_0_0 (CPU device 0)"
I0126 14:16:34.870421 2312 libtorch.cc:2674] "TRITONBACKEND_ModelInstanceFinalize: delete instance state"
terminate called after throwing an instance of 'c10::Error'
what(): Method 'forward' is not defined.
Exception raised from get_method at /tmp/tritonbuild/pytorch/build/include/torch/torch/csrc/jit/api/object.h:111 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits, std::allocator >) + 0x98 (0x7f01bc3de4b8 in /opt/tritonserver/backends/pytorch/libc10.so)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&) + 0xe0 (0x7f01bc38c842 in /opt/tritonserver/backends/pytorch/libc10.so)
frame #2: + 0x3853a (0x7f01bc53753a in /opt/tritonserver/backends/pytorch/libtriton_pytorch.so)
frame #3: + 0x22504 (0x7f01bc521504 in /opt/tritonserver/backends/pytorch/libtriton_pytorch.so)
frame #4: + 0x28670 (0x7f01bc527670 in /opt/tritonserver/backends/pytorch/libtriton_pytorch.so)
frame #5: + 0x288d2 (0x7f01bc5278d2 in /opt/tritonserver/backends/pytorch/libtriton_pytorch.so)
frame #6: TRITONBACKEND_ModelInstanceInitialize + 0x491 (0x7f01bc527db1 in /opt/tritonserver/backends/pytorch/libtriton_pytorch.so)
frame #7: + 0x1acc4f (0x7f01c5917c4f in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #8: + 0x1ade97 (0x7f01c5918e97 in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #9: + 0x190265 (0x7f01c58fb265 in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #10: + 0x1908b6 (0x7f01c58fb8b6 in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #11: + 0x19d22d (0x7f01c590822d in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #12: + 0x99ee8 (0x7f01c4f6bee8 in /usr/lib/x86_64-linux-gnu/libc.so.6)
frame #13: + 0x1869bb (0x7f01c58f19bb in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #14: + 0x197e4a (0x7f01c5902e4a in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #15: + 0x19c67c (0x7f01c590767c in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #16: + 0x29764d (0x7f01c5a0264d in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #17: + 0x29ac2c (0x7f01c5a05c2c in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #18: + 0x3f7272 (0x7f01c5b62272 in /opt/tritonserver/bin/../lib/libtritonserver.so)
frame #19: + 0xdc253 (0x7f01c51d7253 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
frame #20: + 0x94ac3 (0x7f01c4f66ac3 in /usr/lib/x86_64-linux-gnu/libc.so.6)
frame #21: clone + 0x44 (0x7f01c4ff7a04 in /usr/lib/x86_64-linux-gnu/libc.so.6)

Aborted (core dumped)

-> The exported model is working well in other environments, but how can i deploy it here ?
-> Method 'forward' is not defined. <-
🙏

[tanmayv25]

The torch.jit.ignore here seems like the issue to me. Triton's PyTorch backend needs forward function to be able to execute inference.