MCPcopy Create free account
hub / github.com/DeepRec-AI/DeepRec / ComputeSpans

Function ComputeSpans

tensorflow/core/kernels/scale_and_translate_op.cc:188–231  ·  view source on GitHub ↗

Computes the spans for the passed kernel, for a input dimension of length input_size transformed by scale and translate to an output dimension of length output_size. Note that there's no requirement that; output_size = input_size * scale.

Source from the content-addressed store, hash-verified

186// length output_size. Note that there's no requirement that;
187// output_size = input_size * scale.
188Status ComputeSpans(OpKernelContext* context,
189 const functor::SamplingKernelType kernel_type,
190 const int64 output_size, const int64 input_size,
191 const float scale, const float translate,
192 const bool antialias, Spans* spans) {
193 switch (kernel_type) {
194 case functor::Lanczos1Kernel: {
195 return ComputeSpansCore(context, CreateLanczos1Kernel(), output_size,
196 input_size, scale, translate, antialias, spans);
197 }
198 case functor::Lanczos3Kernel: {
199 return ComputeSpansCore(context, CreateLanczos3Kernel(), output_size,
200 input_size, scale, translate, antialias, spans);
201 }
202 case functor::Lanczos5Kernel: {
203 return ComputeSpansCore(context, CreateLanczos5Kernel(), output_size,
204 input_size, scale, translate, antialias, spans);
205 }
206 case functor::GaussianKernel: {
207 return ComputeSpansCore(context, CreateGaussianKernel(), output_size,
208 input_size, scale, translate, antialias, spans);
209 }
210 case functor::BoxKernel: {
211 return ComputeSpansCore(context, CreateBoxKernel(), output_size,
212 input_size, scale, translate, antialias, spans);
213 }
214 case functor::TriangleKernel: {
215 return ComputeSpansCore(context, CreateTriangleKernel(), output_size,
216 input_size, scale, translate, antialias, spans);
217 }
218 case functor::KeysCubicKernel: {
219 return ComputeSpansCore(context, CreateKeysCubicKernel(), output_size,
220 input_size, scale, translate, antialias, spans);
221 }
222 case functor::MitchellCubicKernel: {
223 return ComputeSpansCore(context, CreateMitchellCubicKernel(), output_size,
224 input_size, scale, translate, antialias, spans);
225 }
226 default:
227 return errors::InvalidArgument(Printf("Unrecognized kernel type: %d",
228 static_cast<int>(kernel_type)));
229 }
230 return Status::OK();
231}
232
233// Computes the grad spans for the passed kernel.
234// forward_input_size and forward_output_size are the input and output size from

Callers 2

ComputeGradSpansFunction · 0.85
ComputeMethod · 0.85

Calls 11

ComputeSpansCoreFunction · 0.85
CreateLanczos1KernelFunction · 0.85
CreateLanczos3KernelFunction · 0.85
CreateLanczos5KernelFunction · 0.85
CreateGaussianKernelFunction · 0.85
CreateBoxKernelFunction · 0.85
CreateTriangleKernelFunction · 0.85
CreateKeysCubicKernelFunction · 0.85
InvalidArgumentFunction · 0.85
PrintfFunction · 0.85

Tested by

no test coverage detected