huggingface / nanotron Conditional Complexity

Intro

Conditional complexity (also called cyclomatic complexity) is a term used to measure the complexity of software. The term refers to the number of possible paths through a program function. A higher value ofter means higher maintenance and testing costs (infosecinstitute.com).
Conditional complexity is calculated by counting all conditions in the program that can affect the execution path (e.g. if statement, loops, switches, and/or operators, try and catch blocks...).
Conditional complexity is measured at the unit level (methods, functions...).
Units are classified in four categories based on the measured McCabe index: 1-5 (simple units), 6-10 (medium complex units), 11-25 (complex units), 26+ (very complex units).

Learn more...

Conditional Complexity Overall

There are 1,058 units with 10,690 lines of code in units (48.8% of code).

1 very complex units (215 lines of code)
5 complex units (532 lines of code)
33 medium complex units (1,594 lines of code)
66 simple units (1,899 lines of code)
953 very simple units (6,450 lines of code)

Legend:

51+

26-50

11-25

6-10

1-5

Alternative Visuals

3D view of all units

Conditional Complexity per Extension

51+

26-50

11-25

6-10

1-5

Conditional Complexity per Logical Component

primary logical decomposition

51+

26-50

11-25

6-10

1-5

Most Complex Units

Top 50 most complex units

Unit	# lines	McCabe index	# params
def main() in scripts/scaling_benchmarks.py	215	63	0
def __call__() in src/nanotron/data/clm_collator.py	67	33	3
py::array build_mapping_impl() in src/nanotron/data/nemo_dataset/helpers.cpp	164	30	9
def _get_text() in src/nanotron/data/nemo_dataset/__init__.py	84	29	2
py::array build_blocks_mapping_impl() in src/nanotron/data/nemo_dataset/helpers.cpp	149	29	9
def __call__() in src/nanotron/data/clm_collator.py	68	27	3
def set_logger_verbosity_format() in src/nanotron/logging/base.py	55	25	2
def backward() in src/nanotron/nn/ring_attention_lucidrain.py	96	24	2
def __post_init__() in src/nanotron/config/config.py	72	24	1
def lr_scheduler_builder() in src/nanotron/helpers.py	68	22	3
def forward() in src/nanotron/parallel/pipeline_parallel/block.py	66	22	2
def custom_load_state_dict() in src/nanotron/optim/base.py	44	20	4
def _partition_parameters() in src/nanotron/optim/zero.py	48	20	1
def init_model_randomly() in src/nanotron/models/starcoder2.py	61	20	2
def _update_cos_sin_cache() in src/nanotron/nn/rotary.py	35	17	4
def write_to_csv() in src/nanotron/helpers.py	59	17	4
def compute_global_hidden_layer_metrics() in src/nanotron/metrics_logging.py	34	16	2
def forward() in src/nanotron/nn/rotary.py	27	16	4
def serialize() in src/nanotron/config/utils_config.py	31	16	1
def push_to_wandb() in src/nanotron/eval/upload_to_wandb.py	41	15	6
def __post_init__() in src/nanotron/config/config.py	30	15	1
def forward() in src/nanotron/parallel/tensor_parallel/functional.py	111	15	7
def _init_parallel_groups() in src/nanotron/parallel/context.py	46	15	1
def data_generator() in src/nanotron/data/dataloader.py	91	14	0
def collect_parameter_metrics() in src/nanotron/metrics_logging.py	24	13	2
def process_sft() in src/nanotron/data/sft_processing.py	30	13	3
def save_weights() in src/nanotron/serialize/weights.py	62	13	3
def run() in scripts/log_lighteval_to_wandb.py	49	13	1
def pre_training() in src/nanotron/trainer.py	74	12	3
def run_communication() in src/nanotron/parallel/pipeline_parallel/state.py	26	12	2
def distributed_wait_for_completion() in src/nanotron/s3_checkpoints/s3_mover.py	25	12	2
def prepare_sft_dataset() in src/nanotron/data/sft_processing.py	24	11	5
def _get_dataset_mix() in src/nanotron/data/processing.py	39	11	3
def _build_sample_idx() in src/nanotron/data/nemo_dataset/__init__.py	34	11	7
def end() in src/nanotron/logging/timers.py	24	11	1
def eval_single_checkpoint() in src/nanotron/eval/one_job_runner.py	47	11	2
def _all_gather_params() in src/nanotron/optim/zero.py	29	11	1
def _load_model_checkpoint() in src/nanotron/trainer.py	40	11	2
def backward() in src/nanotron/parallel/tensor_parallel/functional.py	52	11	2
def __getitem__() in src/nanotron/data/tokenized_bytes.py	10	10	2
def __iter__() in src/nanotron/data/samplers.py	28	10	1
def get_train_valid_test_split_() in src/nanotron/data/nemo_dataset/dataset_utils.py	25	10	2
def log_all() in src/nanotron/logging/timers.py	18	10	4
def fp32_accum_hook() in src/nanotron/optim/gradient_accumulator.py	48	10	2
def __post_init__() in src/nanotron/config/models_config.py	33	10	1
def _update_dataloader_based_on_training_stages() in src/nanotron/trainer.py	72	10	3
def pre_save_checkpoint() in src/nanotron/trainer.py	40	10	1
def _post_downloading() in src/nanotron/s3_checkpoints/s3_mover.py	19	10	1
def main() in slurm_launcher.py	45	10	0
def __post_init__() in src/nanotron/data/samplers.py	33	9	1