Fix larget larger (#60985)

co63oc · web-flow · commit 0e1676a769ea · 2024-01-22T10:48:27.000+08:00
diff --git a/paddle/phi/infermeta/binary.cc b/paddle/phi/infermeta/binary.cc
@@ -508,7 +508,7 @@ void ConvInferMeta(const MetaTensor& input,
         dilations[i],
         0,
         phi::errors::InvalidArgument(
-            "The dilation of Op(Conv) should be larget than 0, but received "
+            "The dilation of Op(Conv) should be larger than 0, but received "
             "dilation is %d.",
             dilations[i]));
   }
@@ -550,7 +550,7 @@ void ConvInferMeta(const MetaTensor& input,
         strides[i],
         0,
         phi::errors::InvalidArgument(
-            "The stride of Op(Conv) should be larget than 0, but received "
+            "The stride of Op(Conv) should be larger than 0, but received "
             "stride is %d.",
             strides[i]));
   }
@@ -720,7 +720,7 @@ void ConvTransposeInferMeta(const MetaTensor& x,
         strides[i],
         0,
         errors::InvalidArgument(
-            "The stride of Op(Conv) should be larget than 0, but received "
+            "The stride of Op(Conv) should be larger than 0, but received "
             "stride is %d.",
             strides[i]));
   }
diff --git a/paddle/phi/infermeta/fusion.cc b/paddle/phi/infermeta/fusion.cc
@@ -291,7 +291,7 @@ void Conv1dXPUInferMeta(const MetaTensor& x,
       dilations,
       0,
       phi::errors::InvalidArgument(
-          "The dilation of Op(Conv) should be larget than 0, but received "
+          "The dilation of Op(Conv) should be larger than 0, but received "
           "dilation is %d.",
           dilations));
 
@@ -405,7 +405,7 @@ void Conv2dXPUInferMeta(const MetaTensor& x,
         dilations[i],
         0,
         phi::errors::InvalidArgument(
-            "The dilation of Op(Conv) should be larget than 0, but received "
+            "The dilation of Op(Conv) should be larger than 0, but received "
             "dilation is %d.",
             dilations[i]));
   }
@@ -1870,7 +1870,7 @@ void YoloBoxXPUInferMeta(const MetaTensor& x,
       x_dims[x_dims_size - 1],
       4,
       phi::errors::InvalidArgument(
-          "The last dim of x should be larget than 4, but received "
+          "The last dim of x should be larger than 4, but received "
           " is %d.",
           x_dims[x_dims_size - 1]));
   // compute left out_dims
@@ -1993,7 +1993,7 @@ void ConvTransposeXPUInferMeta(const MetaTensor& x,
         strides[i],
         0,
         errors::InvalidArgument(
-            "The stride of Op(Conv) should be larget than 0, but received "
+            "The stride of Op(Conv) should be larger than 0, but received "
             "stride is %d.",
             strides[i]));
   }
@@ -2331,7 +2331,7 @@ void FusedScaleBiasReluConvBnInferMeta(const MetaTensor& x,
         dilations[i],
         0,
         phi::errors::InvalidArgument(
-            "The dilation of Op(Conv) should be larget than 0, but received "
+            "The dilation of Op(Conv) should be larger than 0, but received "
             "dilation is %d.",
             dilations[i]));
   }
diff --git a/paddle/phi/kernels/cpu/arg_min_max_kernel.cc b/paddle/phi/kernels/cpu/arg_min_max_kernel.cc
@@ -90,7 +90,7 @@ struct VisitDataArgMinMaxFunctor {
   void apply() const {
     dev_ctx.template Alloc<Tout>(out);
 
-    // if flatten, will construct the new dims for the cacluate
+    // if flatten, will construct the new dims for the calculation
     phi::DDim x_dims;
     phi::DDim out_dims;
     int new_axis = axis;
diff --git a/paddle/phi/kernels/cpu/batch_norm_grad_kernel.cc b/paddle/phi/kernels/cpu/batch_norm_grad_kernel.cc
@@ -180,8 +180,8 @@ void BatchNormGradFunctor(const Context& ctx,
     bias_arr.setZero();
   }
 
-  int scale_coefff = use_global_stats ? 1 : N * sample_size;
-  const auto scale_inv_var_nhw = scale_arr * inv_var_arr / scale_coefff;
+  int scale_coeff = use_global_stats ? 1 : N * sample_size;
+  const auto scale_inv_var_nhw = scale_arr * inv_var_arr / scale_coeff;
 
   DenseTensor dy_sum;
   dy_sum.Resize({C});
@@ -213,7 +213,7 @@ void BatchNormGradFunctor(const Context& ctx,
         ConstEigenArrayMap<T> y_data(x.data<T>(), sample_size, N * C);
         for (int nc = 0; nc < N * C; ++nc) {
           x_data.col(nc) = (y_data.col(nc) - bias_arr(nc % C)) /
-                               scale_inv_var_nhw(nc % C) / scale_coefff +
+                               scale_inv_var_nhw(nc % C) / scale_coeff +
                            mean_arr(nc % C);
         }
       }
@@ -261,7 +261,7 @@ void BatchNormGradFunctor(const Context& ctx,
         ConstEigenArrayMap<T> y_data(x.data<T>(), C, N * sample_size);
         for (int nhw = 0; nhw < N * sample_size; nhw++) {
           x_data.col(nhw) =
-              (y_data.col(nhw) - bias_arr) / scale_inv_var_nhw / scale_coefff +
+              (y_data.col(nhw) - bias_arr) / scale_inv_var_nhw / scale_coeff +
               mean_arr;
         }
       }
diff --git a/paddle/phi/kernels/cpu/batch_norm_kernel.cc b/paddle/phi/kernels/cpu/batch_norm_kernel.cc
@@ -70,7 +70,7 @@ void BatchNormKernel(const Context& ctx,
       5,
       phi::errors::InvalidArgument(
           "The size of input X's dimensions should be less than 6."
-          "But received: the size of input X's dimensionss is [%d]",
+          "But received: the size of input X's dimensions is [%d]",
           x_dims.size()));
   const int N = static_cast<int>(x_dims[0]);
   const int C = static_cast<int>(
diff --git a/paddle/phi/kernels/cpu/class_center_sample_kernel.cc b/paddle/phi/kernels/cpu/class_center_sample_kernel.cc
@@ -68,7 +68,7 @@ void ClassCenterSampleKernel(const Context& dev_ctx,
     unique_label.insert(label_ptr[i]);
   }
 
-  // constrcut a lookup table and get sampled_local_class_center
+  // construct a lookup table and get sampled_local_class_center
   std::vector<T> actual_sampled;
   std::map<T, T> new_class_dict;
   T idx = 0;
@@ -112,9 +112,9 @@ void ClassCenterSampleKernel(const Context& dev_ctx,
   }
 
   // remap the input label to sampled class
-  auto* remmaped_label_ptr = dev_ctx.template Alloc<T>(remapped_label);
+  auto* remapped_label_ptr = dev_ctx.template Alloc<T>(remapped_label);
   for (int64_t i = 0; i < numel; ++i) {
-    remmaped_label_ptr[i] = new_class_dict[label_ptr[i]];
+    remapped_label_ptr[i] = new_class_dict[label_ptr[i]];
   }
 }
 
diff --git a/paddle/phi/kernels/cpu/conv_util.h b/paddle/phi/kernels/cpu/conv_util.h
@@ -122,7 +122,7 @@ inline std::vector<int64_t> ComputeOutputShape(
         dilations[i],
         0,
         phi::errors::InvalidArgument(
-            "The dilation of Op(Conv) should be larget than 0, but received "
+            "The dilation of Op(Conv) should be larger than 0, but received "
             "dilation is %d.",
             dilations[i]));
   }
@@ -156,7 +156,7 @@ inline std::vector<int64_t> ComputeOutputShape(
         strides[i],
         0,
         phi::errors::InvalidArgument(
-            "The stride of Op(Conv) should be larget than 0, but received "
+            "The stride of Op(Conv) should be larger than 0, but received "
             "stride is %d.",
             strides[i]));
   }
@@ -169,7 +169,7 @@ inline std::vector<int64_t> ComputeOutputShape(
           "length must be euqal to 2 for Op(Conv). "
           "But received: input's dimension is %d, input's shape is [%s]; "
           "Attr(stride)'s length is %d, Attr(stride) is [%s]; "
-          "difference of input's dimention and Attr(strides)'s length = %u.",
+          "difference of input's dimension and Attr(strides)'s length = %u.",
           in_dims.size(),
           in_dims,
           strides.size(),
diff --git a/paddle/phi/kernels/cpu/cross_entropy_grad_kernel.cc b/paddle/phi/kernels/cpu/cross_entropy_grad_kernel.cc
@@ -48,8 +48,8 @@ void CrossEntropyWithSoftmaxGradCPUKernel(const CPUContext& dev_ctx,
       axis_dim,
       0,
       phi::errors::InvalidArgument(
-          "The axis dimention should be larger than 0, but received "
-          "axis dimention is %d.",
+          "The axis dimension should be larger than 0, but received "
+          "axis dimension is %d.",
           axis_dim));
 
   const int n = phi::funcs::SizeToAxis(axis_v, logit_grad->dims());
diff --git a/paddle/phi/kernels/cpu/cross_entropy_kernel.cc b/paddle/phi/kernels/cpu/cross_entropy_kernel.cc
@@ -40,8 +40,8 @@ void CrossEntropy(const CPUContext& dev_ctx,
       axis_dim,
       0,
       phi::errors::InvalidArgument(
-          "The axis dimention should be larger than 0, but received "
-          "axis dimention is %d.",
+          "The axis dimension should be larger than 0, but received "
+          "axis dimension is %d.",
           axis_dim));
 
   dev_ctx.template Alloc<T>(out);
diff --git a/paddle/phi/kernels/cpu/eig.h b/paddle/phi/kernels/cpu/eig.h
@@ -120,7 +120,7 @@ void TransposeTwoAxis(const DenseTensor& input,
 }
 
 // Apply eig to a batch of matrices, values, vectors and (intermidiate
-// DenseTensor) info are overritten
+// DenseTensor) info are overwritten
 template <typename T, typename Context>
 void LapackEig(DenseTensor* input,
                DenseTensor* values,
diff --git a/paddle/phi/kernels/cpu/masked_select_kernel.cc b/paddle/phi/kernels/cpu/masked_select_kernel.cc
@@ -31,15 +31,15 @@ void MaskedSelectKernel(const Context& dev_ctx,
   auto expanded_size = funcs::MatrixGetBroadcastBatchPortion(
       common::vectorize(x.dims()), common::vectorize(mask.dims()));
 
-  DDim epxand_dims = common::make_ddim(expanded_size);
-  if (mask.dims() != epxand_dims) {
+  DDim expand_dims = common::make_ddim(expanded_size);
+  if (mask.dims() != expand_dims) {
     ExpandKernel<bool, Context>(
         dev_ctx, mask, IntArray(expanded_size), &mask_expand);
   } else {
     mask_expand = mask;
   }
 
-  if (x.dims() != epxand_dims) {
+  if (x.dims() != expand_dims) {
     ExpandKernel<T, Context>(dev_ctx, x, IntArray(expanded_size), &x_expand);
   } else {
     x_expand = x;
diff --git a/paddle/phi/kernels/cpu/mode_kernel.cc b/paddle/phi/kernels/cpu/mode_kernel.cc
@@ -36,7 +36,7 @@ void ModeKernel(const Context& dev_ctx,
                           "The dims of Input(X) should be greater than 0."));
   }
   auto out_dims = out->dims();
-  // axis < 0, cacluate the real axis
+  // axis < 0, calculate the real axis
   if (axis < 0) axis += in_dims.size();
 
   T* output_data = dev_ctx.template Alloc<T>(out);
diff --git a/paddle/phi/kernels/cpu/top_k_kernel.cc b/paddle/phi/kernels/cpu/top_k_kernel.cc
@@ -147,7 +147,7 @@ void TopkKernel(const Context& dev_ctx,
     phi::funcs::set_constant(dev_ctx, indices, static_cast<int64_t>(0));
     return;
   }
-  // axis < 0, cacluate the real axis
+  // axis < 0, calculate the real axis
   if (axis < 0) {
     axis += in_dims.size();
   }
diff --git a/paddle/phi/kernels/gpu/unique_consecutive_functor.h b/paddle/phi/kernels/gpu/unique_consecutive_functor.h
@@ -412,7 +412,7 @@ static void UniqueConsecutiveDimsCUDATensor(const Context& context,
       out_trans.dims().size(), context, out_trans, out, permute);
 }
 
-// functor for processing a multi-dimentional Tensor
+// functor for processing a multi-dimensional Tensor
 template <typename Context, typename InT>
 struct UniqueConsecutiveDimsCUDAFunctor {
   const Context& ctx_;
diff --git a/paddle/phi/kernels/gpu/unique_kernel.cu b/paddle/phi/kernels/gpu/unique_kernel.cu
@@ -543,7 +543,7 @@ struct UniqueFlattendCUDAFunctor {
   }
 };
 
-// functor for processing a multi-dimentional DenseTensor
+// functor for processing a multi-dimensional DenseTensor
 template <typename Context, typename InT>
 struct UniqueDimsCUDAFunctor {
   const Context& ctx_;

Original file line number	Diff line number	Diff line change
`@@ -180,8 +180,8 @@ void BatchNormGradFunctor(const Context& ctx,`
`180`	`180`	`bias_arr.setZero();`
`181`	`181`	`}`
`182`	`182`
`183`		`- int scale_coefff = use_global_stats ? 1 : N * sample_size;`
`184`		`- const auto scale_inv_var_nhw = scale_arr * inv_var_arr / scale_coefff;`
	`183`	`+ int scale_coeff = use_global_stats ? 1 : N * sample_size;`
	`184`	`+ const auto scale_inv_var_nhw = scale_arr * inv_var_arr / scale_coeff;`
`185`	`185`
`186`	`186`	`DenseTensor dy_sum;`
`187`	`187`	`dy_sum.Resize({C});`
`@@ -213,7 +213,7 @@ void BatchNormGradFunctor(const Context& ctx,`
`213`	`213`	`ConstEigenArrayMap<T> y_data(x.data<T>(), sample_size, N * C);`
`214`	`214`	`for (int nc = 0; nc < N * C; ++nc) {`
`215`	`215`	`x_data.col(nc) = (y_data.col(nc) - bias_arr(nc % C)) /`
`216`		`- scale_inv_var_nhw(nc % C) / scale_coefff +`
	`216`	`+ scale_inv_var_nhw(nc % C) / scale_coeff +`
`217`	`217`	`mean_arr(nc % C);`
`218`	`218`	`}`
`219`	`219`	`}`
`@@ -261,7 +261,7 @@ void BatchNormGradFunctor(const Context& ctx,`
`261`	`261`	`ConstEigenArrayMap<T> y_data(x.data<T>(), C, N * sample_size);`
`262`	`262`	`for (int nhw = 0; nhw < N * sample_size; nhw++) {`
`263`	`263`	`x_data.col(nhw) =`
`264`		`- (y_data.col(nhw) - bias_arr) / scale_inv_var_nhw / scale_coefff +`
	`264`	`+ (y_data.col(nhw) - bias_arr) / scale_inv_var_nhw / scale_coeff +`
`265`	`265`	`mean_arr;`
`266`	`266`	`}`
`267`	`267`	`}`
Original file line number	Diff line number	Diff line change
`@@ -68,7 +68,7 @@ void ClassCenterSampleKernel(const Context& dev_ctx,`
`68`	`68`	`unique_label.insert(label_ptr[i]);`
`69`	`69`	`}`
`70`	`70`
`71`		`- // constrcut a lookup table and get sampled_local_class_center`
	`71`	`+ // construct a lookup table and get sampled_local_class_center`
`72`	`72`	`std::vector<T> actual_sampled;`
`73`	`73`	`std::map<T, T> new_class_dict;`
`74`	`74`	`T idx = 0;`
`@@ -112,9 +112,9 @@ void ClassCenterSampleKernel(const Context& dev_ctx,`
`112`	`112`	`}`
`113`	`113`
`114`	`114`	`// remap the input label to sampled class`
`115`		`- auto* remmaped_label_ptr = dev_ctx.template Alloc<T>(remapped_label);`
	`115`	`+ auto* remapped_label_ptr = dev_ctx.template Alloc<T>(remapped_label);`
`116`	`116`	`for (int64_t i = 0; i < numel; ++i) {`
`117`		`- remmaped_label_ptr[i] = new_class_dict[label_ptr[i]];`
	`117`	`+ remapped_label_ptr[i] = new_class_dict[label_ptr[i]];`
`118`	`118`	`}`
`119`	`119`	`}`
`120`	`120`
Original file line number	Diff line number	Diff line change
`@@ -120,7 +120,7 @@ void TransposeTwoAxis(const DenseTensor& input,`
`120`	`120`	`}`
`121`	`121`
`122`	`122`	`// Apply eig to a batch of matrices, values, vectors and (intermidiate`
`123`		`-// DenseTensor) info are overritten`
	`123`	`+// DenseTensor) info are overwritten`
`124`	`124`	`template <typename T, typename Context>`
`125`	`125`	`void LapackEig(DenseTensor* input,`
`126`	`126`	`DenseTensor* values,`