ChildProcessError: Subprocess terminated

Troubleshooting
#1

I get this error while tuning the tf tutorial model. This is the code:

tuning_option = { “tuner”: “xgb”,

"trials": 10,

"early_stopping": 100,

"measure_option": autotvm.measure_option(
    builder=autotvm.LocalBuilder(build_func="default"), runner=runner
),

"tuning_records": "/content/logs/Inception-v1-autotuning.json/content/logs/Inception-v1-autotuning.json" ,

}

tasks = autotvm.task.extract_from_program(mod[“main”], target=target, params=params)

for i, task in enumerate(tasks):

prefix = "[Task %2d/%2d] " % (i + 1, len(tasks))

tuner_obj = XGBTuner(task, loss_type="rank")

tuner_obj.tune(

    n_trial=min(tuning_option["trials"], len(task.config_space)),

    early_stopping=tuning_option["early_stopping"],

    measure_option=tuning_option["measure_option"],

    callbacks=[
        autotvm.callback.progress_bar(tuning_option["trials"], prefix=prefix),
        autotvm.callback.log_to_file(tuning_option["tuning_records"]),
    ],
)

And this is the error:

Schermata 2022-03-02 alle 18.01.32
Schermata 2022-03-02 alle 18.01.32780×315 38.8 KB

I have installed tvm on colab following this https://github.com/sebinthomas/TVM_colab .

Can anyone help me in solving this issue?

Thank you in advance

#2

Probably you forget to install the python setup about tvm:

You can try to enter the directory by cd tvm/python and run python setup.py install

In Colab, you would better do this for each time you use tvm because Colab will forget all the setups after disconnecting.

#3

Hi there, I meet the same problem recently, have you got the solutions? thank you~