jsonnet_evaluate_file issue on branch allennlp-1.0

[vdenberg]

Hi,

When following the instructions in https://github.com/allenai/dont-stop-pretraining/blob/master/DATA_SELECTION.md, upon running the command ""python -m scripts.train --config training_config/vampire.jsonnet --serialization-dir model_logs/vampire-world --environment VAMPIRE --device 0 -o", I get the following error:

overriding model_logs/vampire-world
2021-03-30 17:05:56,684 - INFO - transformers.file_utils - PyTorch version 1.5.1 available.
Something went wrong during jsonnet_evaluate_file, please report this: [json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - invalid literal; last read: 'Z'
Aborted
Traceback (most recent call last):
File "/path/to/python3.7/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/path/to/python3.7/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/path/to//dont-stop-pretraining/vampire/scripts/train.py", line 89, in
main()
File "/path/to/dont-stop-pretraining/vampire/scripts/train.py", line 85, in main
subprocess.run(" ".join(allennlp_command), shell=True, check=True)
File "/path/to/python3.7/subprocess.py", line 468, in run
output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command 'allennlp train --include-package vampire training_config/vampire.jsonnet -s model_logs/vampire-world' returned non-zero exit status 134.

I notice this is the same error as issue #65.

I also notice training works if I use the master branch, so this seems to be an issue related to the branch the DSP DATA_SELECTION.md requires.

Could you perhaps help out with this? Thank you!

[chuyuanli]

Hi, I have the same problem.
I used Allennlp cli to tune with the command allennlp tune config/aa.jsonnet config/hparams.json --serialization-dir result/hpo --study-name test --skip-if-exists.

And I got the error:

/allennlp_optuna/commands/tune.py:55: ExperimentalWarning: AllenNLPExecutor is experimental (supported from v1.4.0). The interface can change in the future.
include_package=include_package,
Something went wrong during jsonnet_evaluate_file, please report this: [json.exception.parse_error.101] parse error at line 1, column 1: syntax error while parsing value - invalid literal; last read: 'a'
Aborted

I don't understand this jsonnet_evaluate_file. Any idea?
Thanks!

[himkt]

Hello @vdenberg, I came here because @chuyuanli mentioned this issue in himkt/allennlp-optuna#33. In the issue of allennlp-optuna, the error occurred when passing a string literal to std.parseJson.

std.parseJson(std.extVar("some_var"))
# std.extVar("some_var") => "hoge"
std.pareJson("hoge") # => json.exception.parse_error!

[dendrenn]

Just to clarify for anyone who may come across this in the future -- replace the following two lines in vampire.jsonnet and the command should work as expected:

# "z_dropout": std.parseJson("Z_DROPOUT"),
"z_dropout": std.extVar("Z_DROPOUT"),

# "lr": std.parseJson("LEARNING_RATE"),
"lr": std.parseJson(std.extVar("LEARNING_RATE")),