if __name__ == "__main__": if len(sys.argv) > 1: algo = Algo() while True: #need to set Beta[accel] to 1 #algo.updateBeta([0,1,0,0]) #algo.updateBetaFinal([0,1,0,0]) var = [2] algo.reset_ygradient() for point in algo.build_gradient(vars = var): algo.updateBeta(point) perf = gameStrat(strat = 6, totalgames=3,Algo=algo, \ renderme = False) y = evalGames(perf) algo.update_ygradient(y) #update algo updateB = algo.eval() algo.updateBetaFinal(updateB) #Exit Training Conditions
algo = Algo() #cheats, init-Betas algo.updateBetaFinal([0, 0, 0, .1]) env = createEnvNoisy() ind = 0 while True: var = [0, 1, 2, 3] ep = [0.2, 0.1, .005, 0.01] algo.reset_ygradient() algo.reset_ymisc() for point in algo.build_gradient(vars=var, eps=ep): algo.updateBeta(point) ret = gameStrat(strat = 6, totalgames=50,Algo=algo, \ renderme = True, env = env) algo.update_ygradient(ret['perf']) algo.update_ymisc(ret['loss']) #update algo updateB = algo.eval(l=0.5, logbestrun=True, logbestrunloss=True) algo.updateBetaFinal(updateB) logit(logMt=(ind, algo))
algo = Algo() #cheats, init-Betas algo.updateBetaFinal([0,0,0,.1]) env = createEnvNoisy() ind = 0 while True: var = [0,1,2,3] ep = [0.2,0.1,.005,0.01] algo.reset_ygradient() algo.reset_ymisc() for point in algo.build_gradient(vars = var, eps = ep): algo.updateBeta(point) ret = gameStrat(strat = 6, totalgames=50,Algo=algo, \ renderme = True, env = env) algo.update_ygradient(ret['perf']) algo.update_ymisc(ret['loss']) #update algo updateB = algo.eval(l = 0.5, logbestrun=True, logbestrunloss=True) algo.updateBetaFinal(updateB)
from learning import Algo a = Algo() ret = a.permute_variables( [[0,0,0,0]], [ (1,(-.1,.1)), (2,(-2,1)) ] ) print 'test: ', str(ret) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars = [1,2]) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars = [1,2], no_origin = True) print 'test: ', str(ret) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars = [1,2,3,4], no_origin = True) print 'testlen: ', str(len(ret))
# options, arguments = p.parse_args() if __name__ == "__main__": if len(sys.argv) > 1: algo = Algo() while True: #need to set Beta[accel] to 1 #algo.updateBeta([0,1,0,0]) #algo.updateBetaFinal([0,1,0,0]) var = [2] algo.reset_ygradient() for point in algo.build_gradient(vars=var): algo.updateBeta(point) perf = gameStrat(strat = 6, totalgames=3,Algo=algo, \ renderme = False) y = evalGames(perf) algo.update_ygradient(y) #update algo updateB = algo.eval() algo.updateBetaFinal(updateB) #Exit Training Conditions
from learning import Algo a = Algo() ret = a.permute_variables([[0, 0, 0, 0]], [(1, (-.1, .1)), (2, (-2, 1))]) print 'test: ', str(ret) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars=[1, 2]) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars=[1, 2], no_origin=True) print 'test: ', str(ret) print 'testlen: ', str(len(ret)) ret = a.build_gradient(vars=[1, 2, 3, 4], no_origin=True) print 'testlen: ', str(len(ret))