Cannot find the 'metadata.csv'

[dremofly]

After run the code, I'm sorry to find 'metadata.csv' does not exist. And I could not find it in the homepage of TGS. I was wondering how to solve this problem.

[jakubczakon]

You need to generate it first by running

neptune run --config configs/neptune.yaml main.py prepare_metadata

remember to specify the paths in the neptune.yaml first.

Thanks for spotting this. The Readme.md was missing this part of the instruction.

[saurabh502]

How to generate metadata.csv using python only?

when I am running below command:
python main.py prepare_metadata

I am getting error:

neptune: Executing in Offline Mode.
neptune: Error: Invalid parameter 'prepare_metadata'. Parameter names must begin with double dash.

[dremofly]

@saurabh502
Right command is
python main.py -- prepare_metadata

[dremofly]

@jakubczakon
Thanks for your replying. I have solved this problem by myself. And I found that there was an error in README. The command of training the model missing a space bettween train and the later --.

[LisburnLad]

@jakubczakon

Hi - could you clarify where the metadata.csv file should be?

I eventually got the prepare_metadata step running in the cloud, which has put the metadata.csv file into SAL-10\output\metadata.csv

I then changed the neptune.yaml file back to "metadata_filepath: /input/metadata.csv" as per the instructions and ran the line to train the network, but this comes up with an error:

"You cannot use input that is not in Neptune storage."

Does the command line to train the network need to include "SAL-10" somewhere? Plus the metadata.csv file is sitting in the output directory of SAL-10 - does it need to move to an input directory?

Thanks

[jakubczakon]

Hi @LisburnLad

There was a typo-like mistake in the command. It should indeed, contain the reference to SAL-10

neptune send 
--worker m-p100 \
--environment pytorch-0.3.1-gpu-py3 \
--config configs/neptune.yaml \
--input /SAL-10/output/metadata.csv \
main.py train --pipeline_name unet

Should work in your case.