My simulation is running to the end successfully. So far I have tried the following submission formats, which I illustrate with example data:
128, 52, 1249
x=128, y=52, z=1249
endx=128, endy=52, endz=1249
All have been rejected. My file is a standard text file with 200,00 entries. What am I missing?
Your file needs to be a .csv file with three columns. The header row should read x, y, z, followed by 200k rows.
Does that answer your question?
Yes, thank you very much, it worked perfectly. Another question: In the notebook, it hints that for dead agents, their coordinates should all (x, y, and z) be set to -1. This seems like a mistake, throwing away information. I have only set the z coordinate to -1 for agents which reach the surface, and then mark them as dead.
Good question, as the wording in the notebook is a bit confusing. The coordinates for dead agents should be their final position and turn. Agents that do not die should have their turn marked as -1 at the end of the simulation. The process you describe is correct.
Can you please clarify the following for the scoring of the submission file for the temporary prediction leader board:
At the end of the simulation:
- Dead Agents --> final coordinates of agent (x,y,z)
- Alive Agents --> final coordinates of agent (x,y,z) OR set x,y,z to -1 (as done in the notebook)?
Hi @bastian.steffens - I believe this was covered in today’s session (recording posted here: Office Hours Q&A for Go with the Flow)
Please let us know if you have any other questions!