huggingface / optimum-quanto 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 355 units with 3,091 lines of code in units (35.9% of code).

0 very long units (0 lines of code)
2 long units (128 lines of code)
20 medium size units (654 lines of code)
66 small units (954 lines of code)
267 very small units (1,355 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 latency() in bench/generation/metrics/latency.py	68	16	7
def parse_quantize_args() in optimum/quanto/subpackage/commands/quantize.py	60	1	1
def main() in bench/generation/evaluate_configurations.py	50	12	0
void mul() in optimum/quanto/library/extensions/cuda/marlin/marlin_cuda.cpp	48	7	10
def main() in bench/generation/evaluate_model.py	46	5	0
def __torch_dispatch__() in optimum/quanto/tensor/weights/qbytes.py	43	14	5
def timing() in bench/kernels/benchmark.py	43	9	3
def from_pretrained() in optimum/quanto/models/diffusers_models.py	42	15	4
def run_benchmark() in bench/kernels/benchmark_w4a16.py	41	13	2
def from_pretrained() in optimum/quanto/models/transformers_models.py	36	14	4
def prepare_inputs_for_generation() in bench/generation/setup/awq.py	33	14	5
def load_from_state_dict() in optimum/quanto/tensor/weights/qbits.py	30	5	8
def run() in optimum/quanto/subpackage/commands/quantize.py	29	6	1
def create() in optimum/quanto/tensor/weights/qbits.py	29	24	9
def unpack_v2() in optimum/quanto/tensor/weights/awq/packed.py	25	2	1
def gen_barchart() in bench/generation/gen_barchart.py	24	5	5
def main() in bench/kernels/benchmark.py	24	7	0
def lib() in optimum/quanto/library/extensions/extension.py	23	6	1
def __torch_dispatch__() in optimum/quanto/tensor/weights/qbits.py	23	10	5
def __init__() in optimum/quanto/tensor/weights/tinygemm/qbits.py	23	8	9
def __torch_dispatch__() in optimum/quanto/tensor/weights/marlin/fp8/packed.py	21	8	5
def qweight() in optimum/quanto/nn/qmodule.py	21	4	1
def pack_weights() in optimum/quanto/tensor/packed.py	20	4	2
def _get_perm() in optimum/quanto/tensor/weights/marlin/int4/packed.py	20	8	0
def __torch_dispatch__() in optimum/quanto/tensor/weights/tinygemm/packed.py	20	8	5
def forward() in optimum/quanto/tensor/weights/awq/qbits.py	20	3	4
def _process_batch() in bench/generation/metrics/perplexity.py	20	8	7
def prediction_accuracy() in bench/generation/metrics/prediction.py	20	4	4
def equal() in optimum/quanto/tensor/qtensor.py	19	11	2
def load_from_state_dict() in optimum/quanto/tensor/weights/qbytes.py	19	5	8
def __torch_dispatch__() in optimum/quanto/tensor/weights/marlin/int4/packed.py	19	6	5
def forward() in optimum/quanto/tensor/weights/tinygemm/qbits.py	19	4	4
def save_bar_chart() in bench/generation/gen_barchart.py	19	4	5
def forward() in optimum/quanto/tensor/weights/marlin/fp8/qbits.py	18	3	4
def get_column_permutation() in optimum/quanto/tensor/weights/marlin/fp8/packed.py	18	2	1
def __torch_dispatch__() in optimum/quanto/tensor/weights/awq/packed.py	18	8	5
def forward() in optimum/quanto/tensor/qbits.py	17	6	2
def setup() in bench/generation/setup/awq.py	17	3	5
def __init__() in optimum/quanto/tensor/weights/marlin/fp8/qbits.py	16	5	8
def __init__() in optimum/quanto/tensor/weights/marlin/int4/qbits.py	16	6	10
def pack_v2() in optimum/quanto/tensor/weights/awq/packed.py	16	4	1
def main() in bench/kernels/benchmark_w4a16.py	16	3	0
def qbytes_mm_impl_cuda() in optimum/quanto/library/qbytes_mm.py	15	8	3
def pack_fp8_as_int32() in optimum/quanto/tensor/weights/marlin/fp8/packed.py	15	4	1
def forward() in optimum/quanto/tensor/weights/marlin/int4/qbits.py	15	3	4
def pack() in optimum/quanto/tensor/weights/marlin/int4/packed.py	15	3	1
def optimize() in optimum/quanto/tensor/weights/qbits.py	15	3	1
def setup() in bench/generation/setup/hqq.py	15	4	5
def __torch_function__() in optimum/quanto/calibrate.py	14	10	4
def qbytes_mm_impl_mps() in optimum/quanto/library/qbytes_mm.py	14	7	3