2 - name: SampLeNet_example
4 # list of launchers for your topology.
6 # launcher framework (e.g. caffe, dlsdk)
8 # device for infer (e.g. for dlsdk cpu, gpu, hetero:cpu,gpu ...)
10 # topology IR (*.prototxt for caffe, *.xml for InferenceEngine, etc)
11 # path to topology is prefixed with directory, specified in "-m/--models" option
12 caffe_model: SampLeNet.prototxt
13 # topology weights binary (*.caffemodel for caffe, *.bin for InferenceEngine)
14 caffe_weights: SampLeNet.caffemodel
15 # launcher returns raw result, so it should be converted
16 # to an appropriate representation with adapter
17 adapter: classification
21 # metrics, preprocessing and postprocessing are typically dataset specific, so dataset field
22 # specifies data and all other steps required to validate topology
23 # there is typically definitions file, which contains options for common datasets and which is merged
24 # during evaluation, but since "sample_dataset" is not used anywhere else, this config contains full definition
26 # uniquely distinguishable name for dataset
27 # note that all other steps are specific for this dataset only
28 # if you need to test topology on multiple datasets, you need to specify
29 # every step explicitly for each dataset
30 - name: sample_dataset
31 # directory where input images are searched.
32 # prefixed with directory specified in "-s/--source" option
33 data_source: sample_dataset/test
34 # parameters for annotation conversion to a common annotation representation format.
35 annotation_conversion:
36 # specified which annotation converter will be used
37 # In order to do this you need to provide your own annotation converter,
38 # i.e. implement BaseFormatConverter interface.
39 # All annotation converters are stored in accuracy_checker/annotation_converters directory.
41 # converter specific parameters.
42 # Full range available options you can find in accuracy_checker/annotation_converters/README.md
43 # relative paths will be merged with "-s/--source" option
44 data_dir: sample_dataset
46 # list of preprocessing, applied to each image during validation
47 # order of entries matters
49 # resize input image to topology input size
50 # you may specify size to which image should be resized
51 # via dst_width, dst_height fields
54 # topology is trained on RGB images, but OpenCV reads in BGR
55 # thence it must be converted to RGB
57 # dataset mean and standard deviation
59 # you may specify precomputed statistics manually or use precomputed values, such as ImageNet as well
60 mean: (125.307, 122.961, 113.8575)
61 std: (51.5865, 50.847, 51.255)
63 # list of metrics, calculated on dataset