You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

Shang Zhizhou 8699f38d08 裁剪transformer模型trt支持；修复tensorRT不支持DeletePass的bug (#28517) ... * skip_layernorm_op done * add unittest * slice op convertor support trt < 6 * skip_layernorm only work in ernie		5 years ago
..
mkldnn	[oneDNN] GRU BF16 kernel (#27731)	5 years ago
CMakeLists.txt	裁剪transformer模型trt支持；修复tensorRT不支持DeletePass的bug (#28517)	5 years ago
conv_fusion_op.cc	…
conv_fusion_op.cu	[bug fix]:Memory increases after adapting the cudnn version to cudnn8 (#27436)	5 years ago
fused_bn_activation_op.cc	…
fused_bn_activation_op.cu	…
fused_bn_activation_op.h	…
fused_bn_add_activation_op.cc	add Fuse bn add act pass (#28196)	5 years ago
fused_bn_add_activation_op.cu	add Fuse bn add act pass (#28196)	5 years ago
fused_bn_add_activation_op.h	add Fuse bn add act pass (#28196)	5 years ago
fused_elemwise_activation_op.cc	Add macro BOOST_GET to enrich the error information of boost :: get (#24175)	5 years ago
fused_elemwise_activation_op.cu	…
fused_elemwise_activation_op.h	refine fused_elemwise_activation error message (#27734)	5 years ago
fused_embedding_eltwise_layernorm_op.cc	…
fused_embedding_eltwise_layernorm_op.cu	…
fused_embedding_fc_lstm_op.cc	fix error msg of fused_embedding_fc_lstm_op, test=develop (#27231)	5 years ago
fused_embedding_fc_lstm_op.h	…
fused_embedding_seq_pool_op.cc	Add macro BOOST_GET to enrich the error information of boost :: get (#24175)	5 years ago
fused_embedding_seq_pool_op.h	Polish the error message of fc, fused_fc_elementwise_layernorm and fused_embedding_seq_pool. (#27692)	5 years ago
fused_fc_elementwise_layernorm_op.cc	Polish some error message in opeators (#27876)	5 years ago
fused_fc_elementwise_layernorm_op.cu	…
fusion_conv_inception_op.cc	…
fusion_conv_inception_op.cu	[CUDNN8 support] : support CUDNN8 (#25664)	5 years ago
fusion_group_op.cc	Add assign to fusion_group and enhance inplace execution in fusion_group. (#26121)	5 years ago
fusion_group_op.cu.cc	…
fusion_group_op.h	Add assign to fusion_group and enhance inplace execution in fusion_group. (#26121)	5 years ago
fusion_group_op_test.cc	Add assign to fusion_group and enhance inplace execution in fusion_group. (#26121)	5 years ago
fusion_gru_op.cc	Quant op dev (#25932)	5 years ago
fusion_gru_op.h	…
fusion_lstm_op.cc	…
fusion_lstm_op.h	…
fusion_repeated_fc_relu_op.cc	…
fusion_repeated_fc_relu_op.h	…
fusion_seqconv_eltadd_relu_op.cc	fix bug sequececonv_eltadd_relu_fuse_pass (#27404)	5 years ago
fusion_seqconv_eltadd_relu_op.h	…
fusion_seqexpand_concat_fc_op.cc	…
fusion_seqexpand_concat_fc_op.h	…
fusion_seqpool_concat_op.cc	…
fusion_seqpool_concat_op.h	…
fusion_seqpool_cvm_concat_op.cc	enhance error messages, test=develop (#27423)	5 years ago
fusion_seqpool_cvm_concat_op.h	…
fusion_squared_mat_sub_op.cc	…
fusion_squared_mat_sub_op.h	…
fusion_transpose_flatten_concat_op.cc	…
fusion_transpose_flatten_concat_op.cu.cc	use iwyu clean include (#27267)	5 years ago
fusion_transpose_flatten_concat_op.h	…
multihead_matmul_op.cc	…
multihead_matmul_op.cu	…
skip_layernorm_op.cc	裁剪transformer模型trt支持；修复tensorRT不支持DeletePass的bug (#28517)	5 years ago
skip_layernorm_op.cu	裁剪transformer模型trt支持；修复tensorRT不支持DeletePass的bug (#28517)	5 years ago