Module caffe2.python.predictor.predictor_py_utils. More...
Functions | |
| def | create_predict_net (predictor_export_meta) |
| def | create_predict_init_net (ws, predictor_export_meta) |
| def | get_comp_name (string, name) |
| def | GetPlan (meta_net_def, key) |
| def | GetPlanOriginal (meta_net_def, key) |
| def | GetBlobs (meta_net_def, key) |
| def | GetNet (meta_net_def, key) |
| def | GetNetOriginal (meta_net_def, key) |
| def | GetApplicationSpecificInfo (meta_net_def, key) |
| def | AddBlobs (meta_net_def, blob_name, blob_def) |
| def | AddPlan (meta_net_def, plan_name, plan_def) |
| def | AddNet (meta_net_def, net_name, net_def) |
Module caffe2.python.predictor.predictor_py_utils.
| def predictor_py_utils.create_predict_init_net | ( | ws, | |
| predictor_export_meta | |||
| ) |
Return an initialization net that zero-fill all the input and output blobs, using the shapes from the provided workspace. This is necessary as there is no shape inference functionality in Caffe2.
Definition at line 24 of file predictor_py_utils.py.
| def predictor_py_utils.create_predict_net | ( | predictor_export_meta | ) |
Return the input prediction net.
Definition at line 11 of file predictor_py_utils.py.
1.8.14