X-Git-Url: https://fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=blobdiff_plain;f=README.txt;h=a4cd46b3ab7feaf839fac3b94e8268eb2d9f01ca;hb=40d0010b6b76304e340ae734cb9814e714b691cc;hp=b5c410975a0408a4930475b9c301b2aab4431680;hpb=27dea0ab0448511236cb344c17162e84359a14ee;p=picoclvr.git diff --git a/README.txt b/README.txt index b5c4109..a4cd46b 100644 --- a/README.txt +++ b/README.txt @@ -1,9 +1,20 @@ -For the stack experiments: + +====================================================================== +For the stack experiment: ./main.py --task=stack -./main.py --task=stack --stack_fraction_values_for_train=0.75 +Takes ~1h10min on a 4090. + +====================================================================== +For the arithmetic expressions experiments + +# 38M parameters / 250k samples + +./main.py --task=expr -./main.py --task=stack --stack_fraction_values_for_train=0.75 --stack_nb_stacks=3 +# 352M parameters / 2.5M samples, reaches 99.80% after 12 epochs, the + learning rate schedule is obviously terrible -Each takes ~1h10min on a 4090. +./main.py --task=expr --nb_blocks=48 --dim_model=1024 --nb_train_samples=2500000 --result_dir=results_expr_48b_d1024_2.5M +======================================================================