Centaurioun/DeepSpeed

master

DeepSpeed

Grade	Name	Complexity	Churn	Issues
A	deepspeed\comm\reduce_op.py		1	2
C	accelerator\cpu_accelerator.py	66	1	29
B	deepspeed\module_inject\containers\features\gated_mlp.py	18	1	15
B	accelerator\npu_accelerator.py	68	1	22
A	.github\workflows\amd-mi100.yml		1	-
A	csrc\cpu\comm\ccl.cpp	40	1	-
A	bin\ds_ssh		1	2
F	deepspeed\runtime\zero\mics.py	50	1	69
B	deepspeed\ops\transformer\inference\triton\softmax.py	8	1	17
B	deepspeed\runtime\zero\mics_utils.py	31	1	24
F	deepspeed\ops\transformer\inference\triton\matmul_ext.py	53	1	151
B	deepspeed\ops\transformer\inference\triton\mlp.py	8	1	15
D	deepspeed\ops\transformer\inference\triton\attention.py	27	1	50
B	deepspeed\runtime\zero\partitioned_param_profiler.py	13	1	13
A	docs\Gemfile		1	-
A	docs\404.html		1	-
B	deepspeed\module_inject\containers\features\hybrid_engine.py	19	1	19
A	csrc\quantization\swizzled_quantize.cu		1	-
A	csrc\quantization\quant_reduce.cu		1	-
A	csrc\aio\py_test\run_read_sweep.sh		1	3