saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-05-14 20:48:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

make sure samples run (#2276)

Browse source 
Adjust instructions for imagenet. Fix compile warning.
 Fix input/output names. Update build/run instructions.
This commit is contained in:
Dmitri Smirnov 2019-10-29 10:07:35 -07:00 committed by GitHub
parent d1096b524f
commit 09eb8ff8b8
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 17 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

6
samples/c_cxx/fns_candy_style_transfer/fns_candy_style_transfer.c
View file

@ -261,16 +261,16 @@ int main(int argc, char* argv[]) {
if (execution_provider)
{
if (tcscmp(execution_provider, ORT_TSTR("cpu"))) {
if (tcscmp(execution_provider, ORT_TSTR("cpu")) == 0) {
// Nothing; this is the default
} else if (tcscmp(execution_provider, ORT_TSTR("cuda"))) {
} else if (tcscmp(execution_provider, ORT_TSTR("cuda")) == 0) {
#ifdef USE_CUDA
enable_cuda(session_options);
#else
puts("CUDA is not enabled in this build.");
return -1;
#endif
} else if (tcscmp(execution_provider, ORT_TSTR("dml"))) {
} else if (tcscmp(execution_provider, ORT_TSTR("dml")) == 0) {
#ifdef USE_DML
enable_dml(session_options);
#else

16
samples/c_cxx/imagenet/README.md
View file

@ -9,7 +9,7 @@ WARNING: If you want to train the model by yourself, you need at least 500GB dis
# Install tensorflow
Install Python 3.x from [python.org](https://www.python.org/), then execute
```
pip install --upgrade tensorflow
pip install --upgrade tensorflow==1.14
```
For more information, see [Install Tensorflow](https://www.tensorflow.org/install)
@ -36,10 +36,12 @@ tar -zxvf inception_v4_2016_09_09.tar.gz
The [Inception V4] zip file only contains a single checkpoint file: inception_v4.ckpt. It can't be directly used for inferencing.
You need to combine the network definition and the checkpoint. Please follow the steps below:
1. Export the graph
Create an new folder. At there, execute
1. Export the graph.
```
git clone https://github.com/tensorflow/models .
git clone https://github.com/tensorflow/models
# Copy inception_v4.ckpt into models
cd models
# Ignore deprecation warnings
python research\slim\export_inference_graph.py --model_name=inception_v4 --output_file=grpah.pb
```
@ -56,6 +58,12 @@ pip install --upgrade tf2onnx
python -m tf2onnx.convert --input inception_v4.pb --inputs input:0 --outputs InceptionV4/Logits/Predictions:0 --opset 10 --output inception_v4.onnx
```
You should see messages like these:
INFO - Successfully converted TensorFlow model inception_v4.pb to ONNX
INFO - ONNX model is saved at inception_v4.onnx
# Run the inferencing
In your build dir of onnxruntime_samples, search for "image_classifier.exe" and run
```

5
samples/c_cxx/imagenet/main.cc
View file

@ -124,11 +124,10 @@ class Validator : public OutputCollector<TCharString> {
VerifyInputOutputCount(session_);
Ort::AllocatorWithDefaultOptions ort_alloc;
{
char* t;
session_.GetInputName(0, ort_alloc);
char* t = session_.GetInputName(0, ort_alloc);
input_name_ = my_strdup(t);
ort_alloc.Free(t);
session_.GetOutputName(0, ort_alloc);
t = session_.GetOutputName(0, ort_alloc);
output_name_ = my_strdup(t);
ort_alloc.Free(t);
}