renom_img.api.inference ¶
-
class
Detector
( url='http://localhost' , port='8080' ) ¶ -
Bases:
object
This class allows you to pull models trained in the ReNomIMG GUI.
Parameters: - url ( string ) – The running ReNomIMG server URL.
- port ( string ) – The running ReNomIMG server port number.
-
model_info
¶ -
This function returns information of pulled model.
Example
>>> from renom_img.api.inference.detector import Detector >>> detector = Detector() >>> detector.pull() >>> print(detector.model_info)
-
predict
( img_list ) ¶ -
Perform prediction for the given image.
Parameters: img_list ( string , list , ndarray ) – Path to the image, list of path or ndarray can be passed. Example
>>> from renom_img.api.inference.detector import Detector >>> detector = Detector() >>> detector.pull() >>> detector.predict(path_to_image) { {'box':[0.2, 0.1, 0.5, 0.3], 'class':0, 'name': 'dog', 'score':0.5} }
-
pull
( ) ¶ -
Pull trained weights from ReNomIMG server. Trained weight will be downloaded into current directory.
Example
>>> from renom_img.api.inference.detector import Detector >>> detector = Detector() >>> detector.pull()
-
class
Classifier
( url='http://localhost' , port='8080' ) ¶ -
Bases:
object
This class allows you to pull trained classification models that are deployed on the ReNomIMG GUI.
Parameters: - url ( string ) – The running ReNomIMG server URL.
- port ( string ) – The running ReNomIMG server port number.
-
model_info
¶ -
This function returns the information of the model you have downloaded.
Example
>>> from renom_img.api.inference.classifier import Classifier >>> classifier = Classifier() >>> classifier.pull() >>> print(classifier.model_info)
-
predict
( img_list , return_scores=False ) ¶ -
Perform prediction for the given image(s).
Parameters: - img_list ( string , list , ndarray ) – Accepts an image path as a string, list of image paths or an image converted to a numpy array.
- return_scores ( boolean ) – Flag to return prediction scores for all classes for each image. Default is False.
Returns: List of predicted classes for each image
Example
>>> from renom_img.api.inference.classifier import Classifier >>> classifier = Classifier() >>> classifier.pull() >>> >>> classifier.predict(path_to_image) 3 # Predicts class index = 3 (index starts from 0) >>> >>> # Return scores for all classes by setting 'return_scores' = True) >>> classifier.predict(path_to_image, return_scores=True) (3, array([[0., 0., 0.0150, 0.9850]], dtype=float32)) # Predicts class index = 3, with score = 98.50% }
-
pull
( ) ¶ -
Download trained weights for the deployed classification model from the ReNomIMG server. The trained weights will be downloaded into the current working directory.
Example
>>> from renom_img.api.inference.classifier import Classifier >>> classifier = Classifier() >>> classifier.pull()
-
class
Segmenter
( url='http://localhost' , port='8080' ) ¶ -
Bases:
object
This class allows you to pull trained segmentation models that are deployed on the ReNomIMG GUI.
Parameters: - url ( string ) – The running ReNomIMG server URL.
- port ( string ) – The running ReNomIMG server port number.
-
model_info
¶ -
This function returns the information of the model you have downloaded.
Example
>>> from renom_img.api.inference.segmenter import Segmenter >>> segmenter = Segmenter() >>> segmenter.pull() >>> print(segmenter.model_info)
-
predict
( img_list ) ¶ -
Perform prediction for the given image(s).
Parameters: img_list ( string , list , ndarray ) – Accepts an image path as a string, list of image paths or an image converted to a numpy array. Returns: List of predicted classes for each image Example
>>> from renom_img.api.inference.segmenter import Segmenter >>> segmenter = Segmenter() >>> segmenter.pull() >>> segmenter.predict(path_to_image) }
-
pull
( ) ¶ -
Download trained weights for the deployed segmentation model from the ReNomIMG server. The trained weights will be downloaded into the current working directory.
Example
>>> from renom_img.api.inference.segmenter import Segmenter >>> segmenter = Segmenter() >>> segmenter.pull()