TVMC tune ONNX2.7 出错

[hxf0223]

TVM/LLVM均使用最新master编译安装。Ubuntu 22.04环境变量如下：

export TVM_HOME=~/work/tvm
export PYTHONPATH=$TVM_HOME/python:${PYTHONPATH}
alias tvmc='python -m tvm.driver.tvmc'

参照 Compiling and Optimizing a Model with TVMC进行compile, run, tune。

#tvmc compile --target "llvm" --input-shapes "data:[1,3,224,224]" -o resnet50-v2-7.tar resnet50-v2-7.onnx
tvmc tune --target="llvm" -o resnet50-v2-7-autotune-records.json resnet50-v2-7.onnx

compie, run均没有错误，tvmc tune时有如下错误，请问需要如何处理呢？

tvmc tune log（路径删除个人home路径）：

tvm/python/tvm/relay/frontend/onnx.py:5948: UserWarning: Input data has unknown dimension shapes: ['N', 3, 224, 224]. Specifying static values may improve performance
  warnings.warn(warning_msg)
/tvm/python/tvm/target/target.py:280: UserWarning: target_host parameter is going to be deprecated. Please pass in tvm.target.Target(target, host=target_host) instead.
  warnings.warn(
Traceback (most recent call last):
  File "/tvm/python/tvm/autotvm/task/task.py", line 523, in _prod_length
    num_iter = int(np.prod([get_const_int(axis.dom.extent) for axis in axes]))
  File "/tvm/python/tvm/autotvm/task/task.py", line 523, in <listcomp>
    num_iter = int(np.prod([get_const_int(axis.dom.extent) for axis in axes]))
  File "/tvm/python/tvm/autotvm/utils.py", line 132, in get_const_int
    raise ValueError("Expect value to be constant int")
ValueError: Expect value to be constant int

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/tvm/python/tvm/autotvm/task/task.py", line 613, in compute_flop
    ret = traverse(sch.outputs)
  File "/tvm/python/tvm/autotvm/task/task.py", line 592, in traverse
    num_element = _prod_length(op.axis)
  File "/tvm/python/tvm/autotvm/task/task.py", line 525, in _prod_length
    raise FlopCalculationError("The length of axis is not constant. ")
tvm.autotvm.task.task.FlopCalculationError: The length of axis is not constant. 

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/anaconda3/lib/python3.9/runpy.py", line 197, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/anaconda3/lib/python3.9/runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "/tvm/python/tvm/driver/tvmc/__main__.py", line 24, in <module>
    tvmc.main.main()
  File "/tvm/python/tvm/driver/tvmc/main.py", line 115, in main
    sys.exit(_main(sys.argv[1:]))
  File "/tvm/python/tvm/driver/tvmc/main.py", line 103, in _main
    return args.func(args)
  File "/tvm/python/tvm/driver/tvmc/autotuner.py", line 273, in drive_tune
    tune_model(
  File "/tvm/python/tvm/driver/tvmc/autotuner.py", line 475, in tune_model
    tasks = autotvm_get_tuning_tasks(
  File "/tvm/python/tvm/driver/tvmc/autotuner.py", line 537, in autotvm_get_tuning_tasks
    tasks = autotvm.task.extract_from_program(
  File "/tvm/python/tvm/autotvm/task/relay_integration.py", line 84, in extract_from_program
    return extract_from_multiple_program([mod], [params], target, ops=ops)
  File "/tvm/python/tvm/autotvm/task/relay_integration.py", line 150, in extract_from_multiple_program
    tsk = create(task_name, args, target=target)
  File "/tvm/python/tvm/autotvm/task/task.py", line 483, in create
    ret.flop = ret.config_space.flop or compute_flop(sch)
  File "/tvm/python/tvm/autotvm/task/task.py", line 615, in compute_flop
    raise RuntimeError(
RuntimeError: FLOP estimator fails for this operator. Error msg: The length of axis is not constant. . Please use `cfg.add_flop` to manually set FLOP for this operator

[E-gethuber]

固定一下onnx的输入输出数据的batch就行。

`import onnx

onnx_model = onnx.load_model(r'./resnet50-v2-7.onnx')
onnx_model.graph.input[0].type.tensor_type.shape.dim[0].dim_value = 1
onnx_model.graph.output[0].type.tensor_type.shape.dim[0].dim_value = 1
onnx.checker.check_model(onnx_model)
onnx.save(onnx_model, r'./resnet50-v2-7-frozen.onnx')
print('frozen model saved.')`

参考：https://discuss.tvm.apache.org/t/error-when-trying-to-tune-the-resnet-model/13431/3