def create_one_train_data(input1, input2): data = dict(in1=input1, in2=input2) q = copy(data) program_set = AdditionProgramSet() addition_env = AdditionEnv(FIELD_ROW, FIELD_WIDTH, FIELD_DEPTH) teacher = AdditionTeacher(program_set) npi_runner = NPIRunner(teacher, recording=True) npi_runner.verbose = True addition_env.reset() run_npi(addition_env, npi_runner, program_set.ADD, data) step_list = {"q": q, "steps": npi_runner.step_list} return step_list
def create_train_data(): program_set = AdditionProgramSet() addition_env = AdditionEnv(FIELD_ROW, FIELD_WIDTH, FIELD_DEPTH) num = 10 questions = create_questions(num) teacher = AdditionTeacher(program_set) npi_runner = NPIRunner(teacher, recording=True) npi_runner.verbose = True steps_list = [] for data in questions: addition_env.reset() q = copy(data) run_npi(addition_env, npi_runner, program_set.ADD, data) steps_list.append({"q": q, "steps": npi_runner.step_list}) return steps_list