• File size measurements show the distribution of size of files.
• Files are classified in four categories based on their size (lines of code): 1-100 (very small files), 101-200 (small files), 201-500 (medium size files), 501-1000 (long files), 1001+(very long files).
• It is a good practice to keep files small. Long files may become "bloaters", code that have increased to such gargantuan proportions that they are hard to work with.

• There are 409 files with 98,155 lines of code.
• 10 very long files (15,415 lines of code)
• 50 long files (32,759 lines of code)
• 105 medium size files (33,306 lines of codeclsfd_ftr_w_mp_ins)
• 67 small files (9,913 lines of code)
• 177 very small files (6,762 lines of code)

There are 100 files with lines longer than 120 characters. In total, there are 244 long lines.

File	# lines	# units	# long lines
fused_bloom_attention_cuda.cu in server/custom_kernels/custom_kernels	219	-	14
fused_attention_cuda.cu in server/custom_kernels/custom_kernels	219	-	14
q_matrix.cu in server/exllamav2_kernels/exllamav2_kernels/cuda	544	-	11
q4_matmul.cu in server/exllama_kernels/exllama_kernels/cuda_func	218	-	10
server.rs in router/src	2105	1	9
mpt_modeling.py in server/text_generation_server/models/custom_modeling	1105	35	8
main.rs in launcher/src	1815	35	6
matrix.cuh in server/exllama_kernels/exllama_kernels	250	-	6
__init__.py in server/text_generation_server/models	1742	2	6
__init__.py in server/text_generation_server/layers/gptq	401	10	6
trtllm.cmake in backends/trtllm/cmake	40	-	5
exllama_ext.cpp in server/exllama_kernels/exllama_kernels	198	3	5
qdq_5.cuh in server/exllamav2_kernels/exllamav2_kernels/cuda/quant	184	-	5
gptq.py in server/text_generation_server/layers/marlin	390	14	4
matrix_view.cuh in server/exllamav2_kernels/exllamav2_kernels/cuda	104	-	4
q_gemm_kernel.cuh in server/exllamav2_kernels/exllamav2_kernels/cuda	507	-	4
ext.cpp in server/exllamav2_kernels/exllamav2_kernels	115	-	4
main.rs in backends/trtllm/src	296	-	3
main.rs in backends/v2/src	203	-	3
flash_phi_moe_modeling.py in backends/gaudi/server/text_generation_server/models/custom_modeling	134	2	3
__init__.py in backends/gaudi/server/text_generation_server/models	984	2	3
__init__.py in backends/gaudi/server/text_generation_server/layers/gptq	371	11	3
backend.rs in backends/v3/src	450	7	3
main.rs in backends/v3/src	217	-	3
lib.rs in router/src	1176	17	3
q4_matrix.cu in server/exllama_kernels/exllama_kernels/cuda_func	166	-	3
flash_phi_moe_modeling.py in server/text_generation_server/models/custom_modeling	134	2	3
qdq_3.cuh in server/exllamav2_kernels/exllamav2_kernels/cuda/quant	146	-	3
q_gemm.cu in server/exllamav2_kernels/exllamav2_kernels/cuda	198	-	3
backend.rs in backends/v2/src	402	7	2
bnb.py in backends/gaudi/server/text_generation_server/layers	93	8	2
mlp.py in backends/gaudi/server/text_generation_server/layers	214	10	2
cli.py in backends/gaudi/server/text_generation_server	297	3	2
queue.rs in backends/v3/src	670	5	2
usage_stats.rs in router/src	369	5	2
flash_cohere_modeling.py in server/text_generation_server/models/custom_modeling	448	15	2
flash_causal_lm.py in server/text_generation_server/models	2009	19	2
bnb.py in server/text_generation_server/layers	93	8	2
w8a8_int.py in server/text_generation_server/layers/compressed_tensors	196	10	2
mlp.py in server/text_generation_server/layers	214	10	2
kv_cache.py in server/text_generation_server/layers/attention	265	9	2
cli.py in server/text_generation_server	301	3	2
util.cuh in server/exllamav2_kernels/exllamav2_kernels/cuda	45	-	2
generation.rs in benchmark/src	186	1	2
nix server.nix in nix	121	-	2
long.js in load_tests	50	2	1
common.js in load_tests	50	2	1
backend.hpp in backends/trtllm/csrc	133	4	1
hardware.hpp in backends/trtllm/csrc	38	2	1
mod.rs in backends/v2/src/client	52	2	1