openai / ppo-ewma - Unit Size

Intro

Unit size measurements show the distribution of size of units of code (methods, functions...).
Units are classified in four categories based on their size (lines of code): 1-20 (small units), 20-50 (medium size units), 51-100 (long units), 101+ (very long units).
You should aim at keeping units small (< 20 lines). Long units may become "bloaters", code that have increased to such gargantuan proportions that they are hard to work with.

Learn more...

Unit Size Overall

There are 250 units with 1,663 lines of code in units (72.7% of code).

0 very long units (0 lines of code)
1 long units (82 lines of code)
15 medium size units (366 lines of code)
29 small units (426 lines of code)
205 very small units (789 lines of code)

Legend:

101+

51-100

21-50

11-20

1-10

Unit Size per Extension

101+

51-100

21-50

11-20

1-10

Unit Size per Logical Component

primary logical decomposition

101+

51-100

21-50

11-20

1-10

Alternative Visuals

3D view of all units

Longest Units

Top 50 longest units

Unit	# lines	McCabe index	# params
def train_fn() in ppo_ewma/train.py	82	8	35
def learn() in ppo_ewma/ppo.py	32	1	2
def read_tb() in ppo_ewma/logger.py	30	9	1
def writekvs() in ppo_ewma/logger.py	27	9	2
def format_model() in ppo_ewma/torch_util.py	26	9	2
def parse_dtype() in ppo_ewma/torch_util.py	26	18	1
def log() in ppo_ewma/log_save_helper.py	26	6	1
def writekvs() in ppo_ewma/logger.py	24	6	2
def dump_kwargs() in ppo_ewma/logger.py	23	3	1
def minibatch_gen() in ppo_ewma/minibatch_optimize.py	22	9	5
def main() in ppo_ewma/train.py	22	2	0
def compute_advantage() in ppo_ewma/ppo.py	22	3	6
def train_with_losses_and_opt() in ppo_ewma/ppo.py	22	6	3
def multi_step() in ppo_ewma/roller.py	22	5	3
def aux_train() in ppo_ewma/ppg.py	21	5	6
def _demo() in ppo_ewma/logger.py	21	2	0
def dumpkvs() in ppo_ewma/logger.py	20	7	1
def singles_to_multi() in ppo_ewma/roller.py	20	11	1
def gather_roller_stats() in ppo_ewma/log_save_helper.py	19	5	2
def make_minibatches() in ppo_ewma/ppg.py	19	6	2
def single_step() in ppo_ewma/roller.py	19	4	2
def process() in ppo_ewma/vec_monitor2.py	18	6	1
def torch_setup() in ppo_ewma/torch_util.py	17	5	2
def save() in ppo_ewma/log_save_helper.py	17	6	1
def forward() in ppo_ewma/ppg.py	17	4	4
def mpi_weighted_mean() in ppo_ewma/logger.py	16	6	2
def profile() in ppo_ewma/logger.py	16	3	1
def compute_losses() in ppo_ewma/ppo.py	16	1	0
def learn() in ppo_ewma/ppg.py	15	1	12
def __init__() in ppo_ewma/logger.py	14	1	2
def make_output_format() in ppo_ewma/logger.py	14	6	3
def tree_multimap() in ppo_ewma/tree_util.py	14	6	3
def explained_variance() in ppo_ewma/torch_util.py	13	3	3
def act() in ppo_ewma/ppg.py	13	1	4
def _fmt_item() in ppo_ewma/minibatch_optimize.py	13	6	2
def forward() in ppo_ewma/impala_cnn.py	13	2	2
def NormedLinear() in ppo_ewma/torch_util.py	12	4	4
def sync_params() in ppo_ewma/torch_util.py	12	4	5
def writekvs() in ppo_ewma/logger.py	12	2	2
def tensor_distr_builder() in ppo_ewma/distr_builder.py	12	3	1
def _get_local_rank_size() in ppo_ewma/torch_util.py	11	3	1
def register_distributions_for_tree_util() in ppo_ewma/torch_util.py	11	1	0
def profile_kv() in ppo_ewma/logger.py	11	3	2
def update() in ppo_ewma/reward_normalizer.py	11	2	2
def walk_pytree() in ppo_ewma/tree_util.py	11	3	3
def __init__() in ppo_ewma/log_save_helper.py	10	1	2
def __call__() in ppo_ewma/log_save_helper.py	10	3	1
def __init__() in ppo_ewma/logger.py	10	2	2
def __init__() in ppo_ewma/impala_cnn.py	10	2	4
def residual() in ppo_ewma/impala_cnn.py	10	6	2