Skip to content

GitLab

"docs/vscode:/vscode.git/clone" did not exist on "2c1811b349f01493fe4976a0fe2c6e4883f098b7"
Commit 93c632cb authored by Shucai Xiao's avatar Shucai Xiao
Browse files

fix bugs in the gpu implementation of the quant_convolution.

parent 585bb331
Hide whitespace changes
Inline Side-by-side
Showing with 6 additions and 5 deletions
src/targets/gpu/quant_convolution.cpp
View file @ 93c632cb
......@@ -34,9 +34,10 @@ argument miopen_quant_convolution::compute(context& ctx,
arg_vec4_x.implicit());
if(status != miopenStatusSuccess)
{
MIGRAPHX_THROW("QUANT_CONVOLUTION: transform input tensfor failed");
MIGRAPHX_THROW("QUANT_CONVOLUTION: transform input tensor failed");
}
// pack input to vec4 format
status = miopenTransformTensor(ctx.get_stream().get_miopen(),
&alpha,
w_desc.get(),
......@@ -46,15 +47,15 @@ argument miopen_quant_convolution::compute(context& ctx,
arg_vec4_w.implicit());
if(status != miopenStatusSuccess)
{
MIGRAPHX_THROW("QUANT_CONVOLUTION: transform weight tensfor failed");
MIGRAPHX_THROW("QUANT_CONVOLUTION: transform weight tensor failed");
}
status = miopenConvolutionForward(ctx.get_stream().get_miopen(),
&alpha,
x_desc.get(),
x_desc_vec4.get(),
arg_vec4_x.implicit(),
w_desc.get(),
args[1].implicit(),
w_desc_vec4.get(),
arg_vec4_w.implicit(),
cd.get(),
algo,
&beta,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment