`fast_lr`

[samarthbhargav]

I'm trying to run the scripts produced by the train_all notebook. The first two work (train_base & train_maml ) work fine, but I get this error when trying to run train_forec.py:

$ python train_forec.py --aug_src_market us --num_epoch 25 --tgt_market uk --batch_size 20 --fast_lr 0.1 --exp_output EVAL/forec-train_all_equal_uk_us_full_aug_ftgt1_fsrc1_shots20.json --data_dir DATA/ --exp_name train_all_equal_uk_us_full_aug_ftgt1_fsrc1 --tgt_fraction 1 --src_fraction 1 --cuda --data_sampling_method equal
usage: FOREC_NeuMF_Engine [-h] [--num_epoch NUM_EPOCH]
                          [--batch_size BATCH_SIZE] [--num_neg NUM_NEG]
                          [--cuda] [--seed SEED] [--exp_name EXP_NAME]
                          [--exp_output EXP_OUTPUT] [--data_dir DATA_DIR]
                          [--tgt_market TGT_MARKET]
                          [--aug_src_market AUG_SRC_MARKET]
                          [--data_sampling_method DATA_SAMPLING_METHOD]
                          [--tgt_fraction TGT_FRACTION]
                          [--src_fraction SRC_FRACTION]
FOREC_NeuMF_Engine: error: unrecognized arguments: --fast_lr 0.1

I'd appreciate any help fixing this!

Thanks!

[hamedrab]

Hi. Thanks for the report. The bug is noted and we will update the code accordingly. We are working on a few other bug fixes as well, so stay tuned for a better version of our code.