Distributed SGD Matrix Factorization (Spark)
How to Run
spark-submit SPARK-PARAMS dsgd_mf.py FACTORS WORKERS ITERATIONS BETA LAMBDA INPUT_PATH W_PATH H_PATH
Each line in the input file is assumed to be i,j,value.
Here is an example:
spark-submit --master local[8] --executor-memory 5G --driver-memory 5G dsgd_mf.py 16 8 100 0.5 1 input.csv w.csv h.csv