From ff4d05a713d859cdad68c375ec24ed71bd3bc4b2 Mon Sep 17 00:00:00 2001 From: li mengyang Date: Fri, 20 Jan 2023 12:27:50 +0800 Subject: [PATCH] fix typo (#4482) --- tools/caffe/caffe.proto | 2 +- tools/mlir/tf_types.cc | 2 +- tools/pnnx/README.md | 8 ++++---- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/tools/caffe/caffe.proto b/tools/caffe/caffe.proto index 5fc52d9e7..95156ecb3 100644 --- a/tools/caffe/caffe.proto +++ b/tools/caffe/caffe.proto @@ -1115,7 +1115,7 @@ message PriorBoxParameter { // Variance for adjusting the prior bboxes. repeated float variance = 6; // By default, we calculate img_height, img_width, step_x, step_y based on - // bottom[0] (feat) and bottom[1] (img). Unless these values are explicitely + // bottom[0] (feat) and bottom[1] (img). Unless these values are explicitly // provided. // Explicitly provide the img_size. optional uint32 img_size = 7; diff --git a/tools/mlir/tf_types.cc b/tools/mlir/tf_types.cc index 381ca869e..92cb16fee 100644 --- a/tools/mlir/tf_types.cc +++ b/tools/mlir/tf_types.cc @@ -293,7 +293,7 @@ bool BroadcastCompatible(TypeRange lhs, TypeRange rhs) // The two types are considered cast compatible if they have dynamically equal // shapes and element type. For element types that do not have subtypes, they // must be equal. However for TensorFlow types such as Resource and Variant, -// that also have subtypes, we recursively check for subtype compatibilty for +// that also have subtypes, we recursively check for subtype compatibility for // Resource types and assume all variant types are cast compatible. If either // one of `a` or `b` have empty subtypes, they are considered cast compatible. // diff --git a/tools/pnnx/README.md b/tools/pnnx/README.md index 78ef0ec3d..7f3640097 100644 --- a/tools/pnnx/README.md +++ b/tools/pnnx/README.md @@ -188,7 +188,7 @@ weight binaries can be listed or modified with any archive application eg. 7zip # PNNX operator PNNX always preserve operators from what PyTorch python api provides. -Here is the netron visualization comparision among ONNX, TorchScript and PNNX with the original PyTorch python code shown. +Here is the netron visualization comparison among ONNX, TorchScript and PNNX with the original PyTorch python code shown. ```python import torch @@ -212,7 +212,7 @@ class Model(nn.Module): # PNNX expression operator PNNX trys to preserve expression from what PyTorch python code writes. -Here is the netron visualization comparision among ONNX, TorchScript and PNNX with the original PyTorch python code shown. +Here is the netron visualization comparison among ONNX, TorchScript and PNNX with the original PyTorch python code shown. ```python import torch @@ -228,7 +228,7 @@ def foo(x, y): # PNNX torch function operator PNNX trys to preserve torch functions and Tensor member functions as one operator from what PyTorch python api provides. -Here is the netron visualization comparision among ONNX, TorchScript and PNNX with the original PyTorch python code shown. +Here is the netron visualization comparison among ONNX, TorchScript and PNNX with the original PyTorch python code shown. ```python import torch @@ -270,7 +270,7 @@ inline module = utils.activations.SiLU pnnx yolov5s.pt inputshape=[1,3,640,640] moduleop=models.common.Focus,models.yolo.Detect ``` -Here is the netron visualization comparision among ONNX, TorchScript and PNNX with the original PyTorch python code shown. +Here is the netron visualization comparison among ONNX, TorchScript and PNNX with the original PyTorch python code shown. ```python import torch