- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I'm running the classification_sample app using SSD_GoogleNetV2 model. Using a 430x430 jpeg image, I get following error: [ ERROR ] Incorrect output dimensions for classification model
Output and image info details are:
[ INFO ] InferenceEngine:
API version ............ 1.0
Build .................. 10478
[ INFO ] Parsing input parameters
[ INFO ] Loading plugin
API version ............ 1.0
Build .................. lnx_20180314
Description ....... MKLDNNPlugin
[ INFO ] Loading network files:
/home/test/cvsdk/models/SSD_GoogleNetV2.xml
/home/test/cvsdk/models/SSD_GoogleNetV2.bin
[ INFO ] Preparing input blobs
[ WARNING ] Image is resized from (430, 430) to (300, 300)
[ INFO ] Batch size is 1
[ INFO ] Preparing output blobs
[ INFO ] Output dims size: 4
[ INFO ] ouputDims[0]: 7, 200
[ ERROR ] Incorrect output dimensions for classification model
API version ............ 1.0
Build .................. 10478
[ INFO ] Parsing input parameters
[ INFO ] Loading plugin
API version ............ 1.0
Build .................. lnx_20180314
Description ....... MKLDNNPlugin
[ INFO ] Loading network files:
/home/test/cvsdk/models/SSD_GoogleNetV2.xml
/home/test/cvsdk/models/SSD_GoogleNetV2.bin
[ INFO ] Preparing input blobs
[ WARNING ] Image is resized from (430, 430) to (300, 300)
[ INFO ] Batch size is 1
[ INFO ] Preparing output blobs
[ INFO ] Output dims size: 4
[ INFO ] ouputDims[0]: 7, 200
[ ERROR ] Incorrect output dimensions for classification model
classification_sample_apptest@nuc:~/cvsdk/samples$ file /home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg
/home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg: JPEG image data, JFIF standard 1.01, resolution (DPI), density 72x72, segment length 16, baseline, precision 8, 430x430, frames 3
test@nuc:~/cvsdk/samples$
/home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg: JPEG image data, JFIF standard 1.01, resolution (DPI), density 72x72, segment length 16, baseline, precision 8, 430x430, frames 3
test@nuc:~/cvsdk/samples$
test@nuc:~/cvsdk/samples$ identify /home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg
/home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg JPEG 430x430 430x430+0+0 8-bit sRGB 108KB 0.000u 0:00.000
test@nuc:~/cvsdk/samples$
/home/test/cvsdk/samples/input_videos/hh-animals-groundhog-4.jpeg JPEG 430x430 430x430+0+0 8-bit sRGB 108KB 0.000u 0:00.000
test@nuc:~/cvsdk/samples$
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Jay07920,
SSD_GoogleNetV2 is a model that outputs bounding boxes, while the sample classification_sample expects labels. This mismatch creates the error you are reporting. To use SSD_GoogleNetV2 , you should look at the samples object_detection_sample_ssd which is working with SSD_GoogleNetV2.
Best regards,
Severine
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page