diff --git a/report/0size_tensor_cpu/error_api.txt b/report/0size_tensor_cpu/error_api.txt
index 809e039a3..f6aabf6b8 100644
--- a/report/0size_tensor_cpu/error_api.txt
+++ b/report/0size_tensor_cpu/error_api.txt
@@ -120,7 +120,6 @@ paddle.multiplex
 paddle.mv
 paddle.nanmedian
 paddle.nn.functional.adaptive_log_softmax_with_loss
-paddle.nn.functional.affine_grid
 paddle.nn.functional.avg_pool3d
 paddle.nn.functional.batch_norm
 paddle.nn.functional.binary_cross_entropy
@@ -131,7 +130,6 @@ paddle.nn.functional.conv2d_transpose
 paddle.nn.functional.conv3d
 paddle.nn.functional.cosine_similarity
 paddle.nn.functional.cross_entropy
-paddle.nn.functional.ctc_loss
 paddle.nn.functional.dice_loss
 paddle.nn.functional.embedding
 paddle.nn.functional.gather_tree
@@ -144,7 +142,6 @@ paddle.nn.functional.kl_div
 paddle.nn.functional.layer_norm
 paddle.nn.functional.linear
 paddle.nn.functional.log_loss
-paddle.nn.functional.margin_cross_entropy
 paddle.nn.functional.max_pool1d
 paddle.nn.functional.max_pool3d
 paddle.nn.functional.max_unpool2d
diff --git a/report/0size_tensor_cpu/error_config.txt b/report/0size_tensor_cpu/error_config.txt
index 31c51a715..fa84e888a 100644
--- a/report/0size_tensor_cpu/error_config.txt
+++ b/report/0size_tensor_cpu/error_config.txt
@@ -5642,11 +5642,6 @@ paddle.nanmedian(Tensor([4, 0],"float64"), keepdim=True, mode="min", )
 paddle.nn.functional.adaptive_log_softmax_with_loss(Tensor([128, 16],"float32"), Tensor([0],"int64"), Tensor([16, 8],"float32"), list[list[Tensor([16, 8],"float32"),Tensor([8, 5],"float32"),],list[Tensor([16, 4],"float32"),Tensor([4, 5],"float32"),],list[Tensor([16, 2],"float32"),Tensor([2, 5],"float32"),],], list[5,10,15,20,], None, )
 paddle.nn.functional.adaptive_log_softmax_with_loss(Tensor([4, 8],"float32"), Tensor([0],"int64"), Tensor([8, 3],"float32"), list[list[Tensor([8, 4],"float32"),Tensor([4, 2],"float32"),],], list[2,4,], None, )
 paddle.nn.functional.adaptive_log_softmax_with_loss(Tensor([8, 8],"float32"), Tensor([0],"int64"), Tensor([8, 3],"float32"), list[list[Tensor([8, 4],"float32"),Tensor([4, 2],"float32"),],], list[2,4,], None, )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float32"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([20, 2, 3],"float32"), Tensor([0],"int64"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=False, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
 paddle.nn.functional.avg_pool3d(Tensor([2, 3, 8, 0, 8],"float64"), kernel_size=list[3,3,3,], stride=list[3,2,1,], padding=list[1,2,1,], ceil_mode=False, exclusive=True, divisor_override=None, data_format="NCDHW", name=None, )
 paddle.nn.functional.avg_pool3d(x=Tensor([2, 3, 8, 0, 8],"float64"), kernel_size=list[3,3,3,], stride=list[3,2,1,], padding=list[1,2,1,], exclusive=True, )
 paddle.nn.functional.batch_norm(Tensor([0, 1, 2, 3],"float32"), Tensor([1],"float32"), Tensor([1],"float32"), Tensor([1],"float32"), Tensor([1],"float32"), training=True, momentum=0.9, epsilon=1e-05, data_format="NCHW", use_global_stats=True, )
@@ -6426,26 +6421,6 @@ paddle.nn.functional.cross_entropy(input=Tensor([2012, 0],"float32"), label=Tens
 paddle.nn.functional.cross_entropy(input=Tensor([2022, 0],"float32"), label=Tensor([2022],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([2036, 0],"float32"), label=Tensor([2036],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([32, 0],"float16"), label=Tensor([32, 1],"int64"), )
-paddle.nn.functional.ctc_loss(Tensor([0, 128, 6625],"float32"), Tensor([128, 25],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), Tensor([2, 3],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 37],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 96],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 3, 15],"float32"), Tensor([3, 2],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([0, 512, 37],"float32"), Tensor([512, 25],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 37],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 96],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 512, 37],"float32"), Tensor([512, 0],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([3, 3, 15],"float32"), Tensor([3, 0],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([40, 128, 6625],"float32"), Tensor([128, 0],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), Tensor([2, 0],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
 paddle.nn.functional.dice_loss(Tensor([0, 2],"float64"), label=Tensor([0, 1],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 0],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 1],"int64"), epsilon=1e-05, )
@@ -8013,18 +7988,6 @@ paddle.nn.functional.linear(x=Tensor([30, 16],"float32"), weight=Tensor([16, 0],
 paddle.nn.functional.linear(x=Tensor([32, 15, 15, 256],"float32"), weight=Tensor([256, 0],"float32"), bias=None, name=None, )
 paddle.nn.functional.linear(x=Tensor([38, 16],"float32"), weight=Tensor([16, 0],"float32"), )
 paddle.nn.functional.log_loss(Tensor([0, 1],"float32"), Tensor([0, 1],"float32"), epsilon=1e-07, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=64.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=0.8, margin2=0.0, margin3=0.15, scale=40.0, group=None, return_softmax=False, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=64.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=32.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float16"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float32"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="sum", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.0, margin3=0.35, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.35, margin2=0.0, margin3=0.0, scale=2.0, group=None, reduction=None, )
 paddle.nn.functional.max_pool1d(Tensor([2, 3, 0],"float64"), 2, 1, list[1,1,], False, False, None, )
 paddle.nn.functional.max_pool1d(x=Tensor([2, 3, 0],"float64"), kernel_size=2, stride=1, padding=list[1,1,], )
 paddle.nn.functional.max_pool3d(Tensor([2, 3, 8, 0, 8],"float64"), kernel_size=list[3,3,3,], stride=list[3,2,1,], padding=list[1,2,1,], return_mask=False, ceil_mode=False, data_format="NCDHW", name=None, )
diff --git a/report/0size_tensor_gpu/error_api.txt b/report/0size_tensor_gpu/error_api.txt
index 71582cfaa..36b64509f 100644
--- a/report/0size_tensor_gpu/error_api.txt
+++ b/report/0size_tensor_gpu/error_api.txt
@@ -180,7 +180,6 @@ paddle.nn.functional.adaptive_log_softmax_with_loss
 paddle.nn.functional.adaptive_max_pool1d
 paddle.nn.functional.adaptive_max_pool2d
 paddle.nn.functional.adaptive_max_pool3d
-paddle.nn.functional.affine_grid
 paddle.nn.functional.avg_pool1d
 paddle.nn.functional.avg_pool2d
 paddle.nn.functional.avg_pool3d
@@ -195,7 +194,6 @@ paddle.nn.functional.conv3d
 paddle.nn.functional.conv3d_transpose
 paddle.nn.functional.cosine_similarity
 paddle.nn.functional.cross_entropy
-paddle.nn.functional.ctc_loss
 paddle.nn.functional.dice_loss
 paddle.nn.functional.embedding
 paddle.nn.functional.gather_tree
@@ -213,7 +211,6 @@ paddle.nn.functional.log_loss
 paddle.nn.functional.log_softmax
 paddle.nn.functional.lp_pool1d
 paddle.nn.functional.lp_pool2d
-paddle.nn.functional.margin_cross_entropy
 paddle.nn.functional.margin_ranking_loss
 paddle.nn.functional.max_pool1d
 paddle.nn.functional.max_pool2d
diff --git a/report/0size_tensor_gpu/error_config.txt b/report/0size_tensor_gpu/error_config.txt
index c82034aed..e598fe686 100644
--- a/report/0size_tensor_gpu/error_config.txt
+++ b/report/0size_tensor_gpu/error_config.txt
@@ -8310,11 +8310,6 @@ paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 3, 5, 7, 7],"float32"), ou
 paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 3, 5, 7, 7],"float32"), output_size=list[2,3,5,], )
 paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 3, 5, 7, 7],"float32"), output_size=list[3,3,3,], )
 paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 3, 5, 7, 7],"float32"), output_size=list[None,3,None,], )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float32"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([20, 2, 3],"float32"), Tensor([0],"int64"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=False, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 1, 120],"float32"), 25, 1, 0, True, False, None, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 3, 32],"float32"), 2, None, 0, True, False, None, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 3, 32],"float32"), kernel_size=2, stride=2, padding="SAME", )
@@ -9288,26 +9283,6 @@ paddle.nn.functional.cross_entropy(input=Tensor([2012, 0],"float32"), label=Tens
 paddle.nn.functional.cross_entropy(input=Tensor([2022, 0],"float32"), label=Tensor([2022],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([2036, 0],"float32"), label=Tensor([2036],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([32, 0],"float16"), label=Tensor([32, 1],"int64"), )
-paddle.nn.functional.ctc_loss(Tensor([0, 128, 6625],"float32"), Tensor([128, 25],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), Tensor([2, 3],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 37],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 96],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 3, 15],"float32"), Tensor([3, 2],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([0, 512, 37],"float32"), Tensor([512, 25],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 37],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 96],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 512, 37],"float32"), Tensor([512, 0],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([3, 3, 15],"float32"), Tensor([3, 0],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([40, 128, 6625],"float32"), Tensor([128, 0],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), Tensor([2, 0],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
 paddle.nn.functional.dice_loss(Tensor([0, 2],"float64"), label=Tensor([0, 1],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 0],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 1],"int64"), epsilon=1e-05, )
@@ -11201,18 +11176,6 @@ paddle.nn.functional.lp_pool1d(Tensor([2, 3, 0],"float32"), 7.0, 2, None, 1, Tru
 paddle.nn.functional.lp_pool1d(Tensor([2, 3, 0],"float32"), norm_type=7, kernel_size=2, stride=2, padding=list[1,], ceil_mode=True, )
 paddle.nn.functional.lp_pool2d(Tensor([0, 3, 32, 32],"float16"), 2, kernel_size=3, stride=2, ceil_mode=False, )
 paddle.nn.functional.lp_pool2d(Tensor([0, 3, 32, 32],"float16"), norm_type=2.0, kernel_size=3, stride=2, padding=0, ceil_mode=False, data_format="NCHW", name=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=64.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=0.8, margin2=0.0, margin3=0.15, scale=40.0, group=None, return_softmax=False, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=64.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=32.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float16"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float32"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="sum", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.0, margin3=0.35, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.35, margin2=0.0, margin3=0.0, scale=2.0, group=None, reduction=None, )
 paddle.nn.functional.margin_ranking_loss(Tensor([0],"float16"), Tensor([0],"float16"), Tensor([0],"float16"), 0.5, "mean", None, )
 paddle.nn.functional.max_pool1d(Tensor([0, 1, 2],"float64"), 2, 2, 0, True, False, None, )
 paddle.nn.functional.max_pool1d(Tensor([0, 3, 16],"float32"), kernel_size=2, stride=2, return_mask=True, )
diff --git a/tester/api_config/7_0_size/0_size_tensor_1_8_1.txt b/tester/api_config/7_0_size/0_size_tensor_1_8_1.txt
index 4fed83a2d..4a74c1d53 100644
--- a/tester/api_config/7_0_size/0_size_tensor_1_8_1.txt
+++ b/tester/api_config/7_0_size/0_size_tensor_1_8_1.txt
@@ -73680,11 +73680,6 @@ paddle.nn.functional.adaptive_max_pool3d(Tensor([0, 1, 5, 5, 5],"float32"), outp
 paddle.nn.functional.adaptive_max_pool3d(Tensor([0, 1, 5, 5, 5],"float64"), output_size=3, return_mask=True, name=None, )
 paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 1, 5, 5, 5],"float32"), output_size=3, return_mask=True, )
 paddle.nn.functional.adaptive_max_pool3d(x=Tensor([0, 1, 5, 5, 5],"float64"), output_size=3, return_mask=True, )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float32"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([1, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([20, 2, 3],"float32"), Tensor([0],"int64"), align_corners=True, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=False, )
-paddle.nn.functional.affine_grid(Tensor([4, 2, 3],"float64"), Tensor([0],"int32"), align_corners=True, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 1, 120],"float32"), 25, 1, 0, True, False, None, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 3, 32],"float32"), 2, None, 0, True, False, None, )
 paddle.nn.functional.avg_pool1d(Tensor([0, 3, 32],"float32"), kernel_size=2, stride=2, padding="SAME", )
@@ -74636,26 +74631,6 @@ paddle.nn.functional.cross_entropy(input=Tensor([2012, 0],"float32"), label=Tens
 paddle.nn.functional.cross_entropy(input=Tensor([2022, 0],"float32"), label=Tensor([2022],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([2036, 0],"float32"), label=Tensor([2036],"int64"), reduction="mean", )
 paddle.nn.functional.cross_entropy(input=Tensor([32, 0],"float16"), label=Tensor([32, 1],"int64"), )
-paddle.nn.functional.ctc_loss(Tensor([0, 128, 6625],"float32"), Tensor([128, 25],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), Tensor([2, 3],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 2, 3],"float64"), labels=Tensor([2, 3],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 37],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 256, 96],"float32"), Tensor([256, 25],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 3, 15],"float32"), Tensor([3, 2],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([0, 4, 34],"float32"), Tensor([4, 4],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([0, 512, 37],"float32"), Tensor([512, 25],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 37],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 256, 96],"float32"), Tensor([256, 0],"int32"), Tensor([256],"int64"), Tensor([256],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([25, 512, 37],"float32"), Tensor([512, 0],"int32"), Tensor([512],"int64"), Tensor([512],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([3, 3, 15],"float32"), Tensor([3, 0],"int32"), Tensor([3],"int64"), Tensor([3],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([4, 4, 34],"float32"), Tensor([4, 0],"int32"), Tensor([4],"int64"), Tensor([4],"int64"), blank=33, reduction="sum", )
-paddle.nn.functional.ctc_loss(Tensor([40, 128, 6625],"float32"), Tensor([128, 0],"int32"), Tensor([128],"int64"), Tensor([128],"int64"), 0, "none", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), Tensor([2, 0],"int32"), Tensor([2],"int64"), Tensor([2],"int64"), 0, "mean", norm_by_times=False, )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="mean", )
-paddle.nn.functional.ctc_loss(Tensor([5, 2, 3],"float64"), labels=Tensor([2, 0],"int32"), input_lengths=Tensor([2],"int64"), label_lengths=Tensor([2],"int64"), blank=0, reduction="none", )
 paddle.nn.functional.dice_loss(Tensor([0, 2],"float64"), label=Tensor([0, 1],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 0],"int64"), epsilon=1e-05, )
 paddle.nn.functional.dice_loss(Tensor([5, 0],"float64"), label=Tensor([5, 1],"int64"), epsilon=1e-05, )
@@ -76462,18 +76437,6 @@ paddle.nn.functional.lp_pool1d(Tensor([2, 3, 0],"float32"), 7.0, 2, None, 1, Tru
 paddle.nn.functional.lp_pool1d(Tensor([2, 3, 0],"float32"), norm_type=7, kernel_size=2, stride=2, padding=list[1,], ceil_mode=True, )
 paddle.nn.functional.lp_pool2d(Tensor([0, 3, 32, 32],"float16"), 2, kernel_size=3, stride=2, ceil_mode=False, )
 paddle.nn.functional.lp_pool2d(Tensor([0, 3, 32, 32],"float16"), norm_type=2.0, kernel_size=3, stride=2, padding=0, ceil_mode=False, data_format="NCHW", name=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=64.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=0.8, margin2=0.0, margin3=0.15, scale=40.0, group=None, return_softmax=False, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=64.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([2, 0],"float64"), label=Tensor([2],"int64"), margin1=1.0, margin2=0.7, margin3=0.2, scale=32.0, group=None, return_softmax=False, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float16"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float32"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="mean", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction="sum", )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, return_softmax=True, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.0, margin3=0.35, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.0, margin2=0.5, margin3=0.0, scale=2.0, group=None, reduction=None, )
-paddle.nn.functional.margin_cross_entropy(Tensor([5, 0],"float64"), Tensor([5],"int64"), return_softmax=False, margin1=1.35, margin2=0.0, margin3=0.0, scale=2.0, group=None, reduction=None, )
 paddle.nn.functional.margin_ranking_loss(Tensor([0],"float16"), Tensor([0],"float16"), Tensor([0],"float16"), 0.5, "mean", None, )
 paddle.nn.functional.max_pool1d(Tensor([0, 1, 2],"float64"), 2, 2, 0, True, False, None, )
 paddle.nn.functional.max_pool1d(Tensor([0, 3, 16],"float32"), kernel_size=2, stride=2, return_mask=True, )