- delete all files in cache/*:
rm cache/*.csv
0b. runpython generate_dre.py
- run
python get_tickers.py
to get the latest tickers, and put them incovarianceforecasting.py
- run example_submission:
python examples_m6/example_m6_entry.py
- upload those files to Kaggle (precise-cache):
kaggle datasets version -p cache -m '2ndround'
- before the saturday when the period ends, run:
python submit_10_notebooks.py
, 3 times (modifying the ports index:PORTS[0]
, thenPORTS[1]
, thenPORTS[2]
) check status with:for i in $(seq 0 22); do kaggle kernels status marcogorelli/f-999-port-1-$i; done;
- repeat steps 0-3 from above!
- update yfinance cache:
kaggle kernels push -p yfinance-data
- check cv-results:
kaggle kernels push -p cv-results
(update the kernel_metadata sources if notebook inputs have changed) make sure to remove the TODO line if it's there! 6a. runkaggle kernels push -p check-naive-cv
6b. to get the best candidates, run the notebookkaggle kernels push -p find-best-from-cv
NOTE: you may want to tighten or loosen the conditions to find final_candidates - run the combine notebook:
kaggle kernels push -p combine
- run the submit notebook:
kaggle kernels push -p submit
- final check:
kaggle kernels push -p final-check
- submit the output of the submit notebook