aws / sagemaker-hyperpod-training-adapter-for-nemo

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 408 units with 3,968 lines of code in units (63.8% of code).

1 very complex units (170 lines of code)
0 complex units (0 lines of code)
10 medium complex units (365 lines of code)
26 simple units (610 lines of code)
371 very simple units (2,823 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 exp_manager() in src/hyperpod_nemo_adapter/utils/exp_manager.py	170	54	3
def torch_call() in src/hyperpod_nemo_adapter/utils/dpo_utils.py	31	21	5
def after_model_validations() in src/hyperpod_nemo_adapter/conf/config_schemas.py	23	17	1
def _should_save_sharded() in src/hyperpod_nemo_adapter/utils/callbacks/checkpoint.py	30	17	3
def get_transformer_layer() in src/hyperpod_nemo_adapter/utils/fsdp_utils.py	37	17	4
def setup() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	48	16	3
def _setup_model() in src/hyperpod_nemo_adapter/collections/parts/fsdp_strategy.py	87	16	2
def before_model_validations() in src/hyperpod_nemo_adapter/conf/config_schemas.py	15	13	2
def _get_max_steps() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	28	13	1
def _get_data_format() in src/hyperpod_nemo_adapter/collections/data/datasets/hf_dataset.py	17	12	2
def main() in src/hyperpod_nemo_adapter/scripts/fp8_cast_bf16.py	49	11	2
def tokenize_dialogs() in src/hyperpod_nemo_adapter/collections/data/vision_dataset.py	24	10	3
def flush_left() in src/hyperpod_nemo_adapter/utils/dpo_utils.py	17	9	2
def _build_model_from_pretrain_peft() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	31	9	2
def _create_data_module() in src/hyperpod_nemo_adapter/collections/parts/sagemaker_trainer_builder.py	13	9	2
def is_global_rank_zero() in src/hyperpod_nemo_adapter/utils/get_rank.py	10	8	0
def get_auto_wrap_policy() in src/hyperpod_nemo_adapter/utils/fsdp_utils.py	38	8	3
def concatenated_forward() in src/hyperpod_nemo_adapter/utils/dpo_utils.py	30	8	5
def pad() in src/hyperpod_nemo_adapter/utils/dpo_utils.py	13	8	3
def get_dataloader() in src/hyperpod_nemo_adapter/collections/data/hf_image_data_module.py	30	8	2
def __init__() in src/hyperpod_nemo_adapter/collections/model/nlp/custom_models/modeling_deepseek.py	33	8	2
def moe_infer() in src/hyperpod_nemo_adapter/collections/model/nlp/custom_models/modeling_deepseek.py	61	8	4
def _transform() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	27	8	2
def _training_strategy() in src/hyperpod_nemo_adapter/collections/parts/sagemaker_trainer_builder.py	8	8	1
def get_interval() in src/hyperpod_nemo_adapter/utils/callbacks/checkpoint.py	37	7	3
def exit_routine() in src/hyperpod_nemo_adapter/utils/tracer_utils.py	13	7	1
def _init_rope() in src/hyperpod_nemo_adapter/collections/model/nlp/custom_models/modeling_deepseek.py	45	7	1
def get_lora_config() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	16	7	1
def save_peft_model() in src/hyperpod_nemo_adapter/collections/parts/fsdp_strategy.py	27	7	2
def after_model_validations() in src/hyperpod_nemo_adapter/conf/config_schemas.py	13	6	1
def _validate_model_type() in src/hyperpod_nemo_adapter/utils/config_utils.py	10	6	2
def get_dataloader_kwargs() in src/hyperpod_nemo_adapter/collections/data/vision_dataset.py	32	6	4
def _get_model_configurable_dict() in src/hyperpod_nemo_adapter/collections/model/sagemaker_base_model.py	12	6	1
def lightning_module_state_dict() in src/hyperpod_nemo_adapter/collections/parts/fsdp_strategy.py	14	6	1
def optimizer_state() in src/hyperpod_nemo_adapter/collections/parts/fsdp_strategy.py	12	6	2
def _create_checkpoint_callbacks() in src/hyperpod_nemo_adapter/collections/parts/sagemaker_trainer_builder.py	24	6	1
def patched_LFA2__init__() in src/hyperpod_nemo_adapter/patches/patch_llama_flash_attn_cp.py	20	6	3
def get_model_validator() in src/hyperpod_nemo_adapter/conf/config_schemas.py	15	5	2
def list_checkpoint_dirs() in src/hyperpod_nemo_adapter/utils/callbacks/local_ckpt_io.py	13	5	1
def load_step() in src/hyperpod_nemo_adapter/utils/callbacks/local_ckpt_io.py	15	5	2
def load_local_metadata() in src/hyperpod_nemo_adapter/utils/callbacks/local_ckpt_io.py	15	5	2
def __init__() in src/hyperpod_nemo_adapter/utils/callbacks/checkpoint.py	25	5	4
def _monitor_candidates() in src/hyperpod_nemo_adapter/utils/callbacks/checkpoint.py	7	5	2
def _save_checkpoints() in src/hyperpod_nemo_adapter/utils/callbacks/checkpoint.py	12	5	2
def get_hf_config_from_name_or_path() in src/hyperpod_nemo_adapter/utils/config_utils.py	11	5	1
def mm_collate_fn() in src/hyperpod_nemo_adapter/collections/data/hf_data_module.py	9	5	1
def get_vocab_size() in src/hyperpod_nemo_adapter/collections/data/dummy_data_module.py	10	5	1
def get_vocab_size() in src/hyperpod_nemo_adapter/collections/data/dummy_dpo_data_module.py	10	5	1
def __call__() in src/hyperpod_nemo_adapter/collections/data/vision_dataset.py	25	5	2
def __iter__() in src/hyperpod_nemo_adapter/collections/data/vision_dataset.py	9	5	1