zzy34407230
/
mindspore2022
Not watched
1
0
Fork 0
Code
Releases 22 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
12207 Commits
18 Branches
584 MB
239 Download
 
 
 
 
 
 
Tree: bc738d4ec3
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from 'bc738d4ec3'
${ noResults }
mindspore2022/mindspore/lite/nnacl
mindspore-ci-bot 5fa1047163 !7688 clear warning of compiler 5 years ago
..
assembly [MSLITE] Optimize depthwise conv 3x3 arm64 5 years ago
fp16 !7648 [MSLITE] deconv winograd fp16 neon 5 years ago
fp32 clear warning of complier 5 years ago
fp32_grad Adam optimizer + Bug Fix 5 years ago
int8 !7712 rename and change lib c++_static to c++_shared 5 years ago
optimize [MSLITE][Develop] modify optimize.so to sdot and fp16 so 5 years ago
quantization Fix bug of leakyRelu_Int8 operator. 5 years ago
CMakeLists.txt fix mindspore-static lib generation and fix cast infershape 5 years ago
README.md update README 5 years ago
arg_min_max.c optimize argmax 5 years ago
arg_min_max.h move nnacl to lite/ 5 years ago
arg_min_max_parameter.h move nnacl to lite/ 5 years ago
arithmetic_common.c move nnacl to lite/ 5 years ago
arithmetic_common.h fix bug of cpu op arithmetic & softmax 5 years ago
arithmetic_parameter.h move nnacl to lite/ 5 years ago
arithmetic_self_parameter.h move nnacl to lite/ 5 years ago
batch_to_space.c move nnacl to lite/ 5 years ago
batch_to_space.h move nnacl to lite/ 5 years ago
batchnorm_parameter.h tod add networks and ops 5 years ago
clip.c Add dropout pass, add clip pass, and fix the fusion problem of transpose 5 years ago
clip.h Add dropout pass, add clip pass, and fix the fusion problem of transpose 5 years ago
common_func.c optimization for fp32 matmul kernel on arm64 5 years ago
common_func.h optimization for fp32 matmul kernel on arm64 5 years ago
concat_parameter.h fix concat memory leak 5 years ago
conv_parameter.h [MSLITE][Develop]deconv winograd input pack and output bias 5 years ago
crop_parameter.h move nnacl to lite/ 5 years ago
depth_to_space.c move nnacl to lite/ 5 years ago
depth_to_space.h move nnacl to lite/ 5 years ago
depth_to_space_parameter.h move nnacl to lite/ 5 years ago
detection_post_process_parameter.h benchmark support multi outputs node & detection_post_process use stable sort & add model detect.tflite 5 years ago
errorcode.h fp16 deconv winograd 5 years ago
flatten.c move nnacl to lite/ 5 years ago
flatten.h move nnacl to lite/ 5 years ago
gather_parameter.h add op_gather_int8 and testcase 5 years ago
instance_norm_parameter.h add lite ops of instanceNorm 5 years ago
l2_norm.c l2norm multi thread and trailing axis 5 years ago
l2_norm.h l2norm multi thread and trailing axis 5 years ago
l2_norm_parameter.h l2norm multi thread and trailing axis 5 years ago
layer_norm_parameter.h add layer_norm op and layer_norm fusion 5 years ago
lsh_projection_parameter.h fix_lsh_projection 5 years ago
matmul_parameter.h [MSLITE][Develop] fix bug of arm cpu matmul: init const tensor in Init function bofore check infershape done 5 years ago
minimal_filtering_generator.c [MSLITE] deconv optimize fp32 5 years ago
minimal_filtering_generator.h [MSLITE] deconv optimize fp32 5 years ago
mul_parameter.h move nnacl to lite/ 5 years ago
nnacl_utils.c move nnacl to lite/ 5 years ago
nnacl_utils.h move nnacl to lite/ 5 years ago
non_max_suppression_parameter.h add op NonMaxSuppression 5 years ago
op_base.h syc op_base act type with schema 5 years ago
optimized_kernel.h change log_adapter 5 years ago
pack.c 1.change int8 conv which not support sdot scheme (using matmul) 5 years ago
pack.h 1.change int8 conv which not support sdot scheme (using matmul) 5 years ago
pad_parameter.h pad reflect and stridedslice infershape 5 years ago
pooling_parameter.h solve average pooling avgMode problem 5 years ago
power.c move nnacl to lite/ 5 years ago
power.h move nnacl to lite/ 5 years ago
power_parameter.h move nnacl to lite/ 5 years ago
predict_parameter.h add predict kernel 5 years ago
prelu_parameter.h optimize prule 5 years ago
prior_box.c move nnacl to lite/ 5 years ago
prior_box.h move nnacl to lite/ 5 years ago
reduce_parameter.h move infershape from internal to nnacl 5 years ago
reshape.c move nnacl to lite/ 5 years ago
reshape.h move nnacl to lite/ 5 years ago
reshape_parameter.h move nnacl to lite/ 5 years ago
resize_parameter.h move nnacl to lite/ 5 years ago
reverse_sequence.c reverse_seqence seq_lengths support int64 5 years ago
reverse_sequence.h reverse_seqence seq_lengths support int64 5 years ago
scale.h scale add relu and relu6 for fusion and calculating 5 years ago
scatter_nd.c move nnacl to lite/ 5 years ago
scatter_nd.h move nnacl to lite/ 5 years ago
shape.h move nnacl to lite/ 5 years ago
sigmoid_parameter.h move nnacl to lite/ 5 years ago
slice_parameter.h fix slice parallel run bug 5 years ago
softmax_parameter.h fix bug of cpu op arithmetic & softmax 5 years ago
sparse_to_dense_parameter.h Add index validity check and adapt for more conditions of sparse_to_dense ops. 5 years ago
split.c Add the function that split size can be -1 at the end of split ops 5 years ago
split.h move nnacl to lite/ 5 years ago
split_parameter.h Remove cast operator which converts data from fp32 to fp16 in models generated by MindSpore. 5 years ago
squeeze.c move nnacl to lite/ 5 years ago
squeeze.h move nnacl to lite/ 5 years ago
squeeze_parameter.h move nnacl to lite/ 5 years ago
stack_parameter.h scale stack fp16 5 years ago
strided_slice.c multi input multi dimension 5 years ago
strided_slice.h support neg stride 5 years ago
transpose.c Add implementation of transpose to adapt different dimensions. 5 years ago
transpose.h Add implementation of transpose to adapt different dimensions. 5 years ago
unsqueeze_parameter.h move nnacl to lite/ 5 years ago
unstack.c move nnacl to lite/ 5 years ago
unstack.h move nnacl to lite/ 5 years ago
where.c move nnacl to lite/ 5 years ago
where.h move nnacl to lite/ 5 years ago
winograd_transform.c optimization for fp32 winograd on arm32 5 years ago
winograd_transform.h replace 8x8 block with 12x8 block in common conv 5 years ago
winograd_utils.c optimize winograd input transform func 5 years ago
winograd_utils.h optimize winograd input transform func 5 years ago
zeroslike.c move nnacl to lite/ 5 years ago
zeroslike.h move nnacl to lite/ 5 years ago

README.md

NNACL(neural network accelerated computing library) is a high performance library of neural network inference computing kernels for ARM.

Delete task

Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.

cancel
confirm

阿对对队

C++ Python Text C Unity3D Asset other