Here you can find the list of models available.

Image Classification

---

Model: NASNetLarge

Category: Image Classification

Class: readyml.imageclassification.NASNetLarge

Reference: Learning Transferable Architectures for Scalable Image Recognition (CVPR 2018)

Example of use:

from readyml import imageclassification as ric
import PIL.Image as Image

## Read an image
image_pil = Image.open("../images/greek_street.jpeg")

## Instantiate the model class
nasnetlarge = ric.NASNetLarge()
## Get categories with a confidence score equal or above 30%
results = nasnetlarge.infer(image_pil, threshold=30)
print(results)

Results: The labels and its percent accuracy.

[
    {
        "label": "monastery",
        "score": 38.63
    }
]

---

Model: MobileNetV2

Category: Image Classification

Class: readyml.imageclassification.MobileNetV2

---

Model: InceptionV3

Category: Image Classification

Class: readyml.imageclassification.InceptionV3

---

Model: Resnet50

Category: Image Classification

Class: readyml.imageclassification.Resnet50

---

Model: Resnet152x4

Category: Image Classification

Class: readyml.imageclassification.Resnet152x4

---

Object Detection

---

Model: Hourglass 512x512

Category: Object Detection

Class: readyml.objectdetection.HourGlass_512x512

Reference:

Example of use:

from readyml import objectdetection as rod
import PIL.Image as Image

## Read an image
image_pil = Image.open("./images/trafalgar.jpg")

## Instantiate the model class
model = rod.HourGlass_512x512()

preds, image = model.infer(image_pil)
im = Image.fromarray(image)
im.save("trafalgar.jpg")

print(preds)

Results: An array of found objects, with the object's label, score, and bounding box coordinates.

[
    {
        "box": [
            1111,
            2219,
            570,
            3696
        ],
        "label": "person",
        "score": 96.78
    },
    {
        "box": [
            2134,
            2842,
            925,
            3666
        ],
        "label": "person",
        "score": 85.62
    }
]

---

Model: Hourglass 1024x1024

Category: Object Detection

Class: readyml.objectdetection.HourGlass_1024x1024

---

Model: Resnet50 v1 fpn 512x512

Category: Object Detection

Class: readyml.objectdetection.Resnet50v1Fpn_512x512

---

Model: Resnet101 v1 fpn 512x512

Category: Object Detection

Class: readyml.objectdetection.Resnet101v1Fpn_512x512

---

Model:

Category: Object Detection

Class: readyml.objectdetection.

---

Model: Resnet50 v2 512x512

Category: Object Detection

Class: readyml.objectdetection.Resnet50v2_512x512

---

Model: Efficientdet D0

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD0

---

Model: Efficientdet D1

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD1

---

Model: Efficientdet D2

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD2

---

Model: Efficientdet D3

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD3

---

Model: Efficientdet D4

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD4

---

Model: Efficientdet D5

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD5

---

Model: Efficientdet D6

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD6

---

Model: Efficientdet D7

Category: Object Detection

Class: readyml.objectdetection.EfficientdetD7

---

Model: SsdMobilenet v2

Category: Object Detection

Class: readyml.objectdetection.SsdMobilenetv2

---

Model: SsdMobilenet v1 Fpn 640x640

Category: Object Detection

Class: readyml.objectdetection.SsdMobilenetv1Fpn_640x640

---

Model: SsdMobilenet v2 Fpn Lite 320x320

Category: Object Detection

Class: readyml.objectdetection.SsdMobilenetv2FpnLite_320x320

---

Model: Resnet50 v1 Fpn 640x640

Category: Object Detection

Class: readyml.objectdetection.Resnet50V1Fpn_640x640

---

Model: Resnet50 v1 Fpn 1024x1024

Category: Object Detection

Class: readyml.objectdetection.Resnet50v1Fpn_1024x1024

---

Model: Resnet101 v1 Fpn 640x640

Category: Object Detection

Class: readyml.objectdetection.Resnet101v1Fpn_640x640

---

Model: Resnet101 v1 Fpn 1024x1024

Category: Object Detection

Class: readyml.objectdetection.Resnet101v1Fpn_1024x1024

---

Model: Resnet152 v1 Fpn 640x640

Category: Object Detection

Class: readyml.objectdetection.Resnet152v1Fpn_640x640

---

Model: Resnet152 v1 Fpn 1024x1024

Category: Object Detection

Class: readyml.objectdetection.Resnet152v1Fpn_1024x1024

---

Model: FasterRcnn Resnet50 v1 640x640

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet50v1_640x640

---

Model: FasterRcnn Resnet50 v1 1024x1024

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet50v1_1024x1024

---

Model: FasterRcnn Resnet50 v1 800x1333

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet50v1_800x1333

---

Model: FasterRcnn Resnet101 v1 640x640

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet101v1_640x640

---

Model: FasterRcnn Resnet101 v1 1024x1024

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet101v1_1024x1024

---

Model: FasterRcnn Resnet101 v1 800x1333

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet101v1_800x1333

---

Model: FasterRcnn Resnet152 v1 640x640

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet152v1_640x640

---

Model: FasterRcnn Resnet152 v1 1024x1024

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet152v1_1024x1024

---

Model: FasterRcnn Resnet152 v1 800x1333

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnResnet152v1_800x1333

---

Model: FasterRcnn Inception Resnetv2 640x640

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnInceptionResnetv2_640x640

---

Model: FasterRcnn Inception Resnetv2 1024x1024

Category: Object Detection

Class: readyml.objectdetection.FasterRcnnInceptionResnetv2_1024x1024

---

Model: MaskRcnn Inception Resnet v2 1024x1024

Category: Object Detection

Class: readyml.objectdetection.MaskRcnnInceptionResnetv2_1024x1024

---

Image Generation

---

Model: BigGanDeep 128

Category: Image Generation

Class: readyml.objectdetection.BigGanDeep128

Example of use:

from readyml import imagegeneration as rig
import PIL.Image as Image

category = 356

model = rig.BigGanDeep128()

new_image = model.infer(category)

im = Image.fromarray(new_image[0])
im.save("myimage.jpeg")

Result: A generated image

---

Model: BigGanDeep 256

Category: Image Generation

Class: readyml.objectdetection.BigGanDeep256

---

Model: BigGanDeep 512

Category: Image Generation

Class: readyml.objectdetection.BigGanDeep512

---

Model: BigGan 128

Category: Image Generation

Class: readyml.objectdetection.BigGan128

---

Model: BigGan 256

Category: Image Generation

Class: readyml.objectdetection.BigGan256

---

Model: BigGan 512

Category: Image Generation

Class: readyml.objectdetection.BigGan512

---

Face Generation

---

Model: Progan 128

Category: Face Generation

Class: readyml.facegeneration.FaceGeneration

Example of use:

from readyml import facegeneration as rfg
import PIL.Image as Image

model = rfg.FaceGeneration()

image = model.infer(num_samples=30)[0]

image = Image.fromarray(image.numpy())
image.save("myimage.jpeg")

Result: A generated face image

---

Face Detection

---

Model: Light Face Detection

Category: Face Detection

Class: readyml.facedetection.FaceDetection

Reference: https://github.com/borhanMorphy/light-face-detection

Example of use:

from readyml import facedetection as rfd
import imageio

img = imageio.imread("../images/faces.jpg")[:,:,:3]

model = rfd.FaceDetectionModel()
preds = model.infer(img)
print(preds)

Result: Bouding boxes coordinates and confidence scores

[
    {
        "box": [
            46,
            19,
            88,
            70
        ],
        "score": 100.0
    },
    {
        "box": [
            94,
            95,
            133,
            146
        ],
        "score": 100.0
    },
    {
        "box": [
            207,
            44,
            253,
            106
        ],
        "score": 100.0
    }
]

---

Pose Detection

---

Model: Movenet Singlepose Lightning

Category: Pose Detection

Class: readyml.posedetection.MovenetSingleposeLightning

Example of use:

from readyml import posedetection as rpd
import PIL.Image as Image

## Read an image
image_pil = Image.open("../images/movenet-singlepose-lightning.jpg")

## Instantiate the model class
model = rpd.MovenetSingleposeLightning()

keypoint_with_scores = model.infer(image_pil)

new_image = model.draw(image_pil, keypoint_with_scores)
im = Image.fromarray(new_image)
im.save("pose-detection.jpeg")

print(keypoint_with_scores)

Result: The original image with the keypoints

---

Image Super Resolution

---

Model: Enhanced Super Resolution GAN

Category: Image Super Resolution

Class: readyml.imagerestoration.MIRNet

Example of use:

from readyml import superresolution as rsr
import PIL.Image as Image
import tensorflow as tf

# Read an image
image_pil = Image.open("../images/lowres.jpg")

# Instantiate the model class
model = rsr.ESRgan()

image = model.infer(image_pil)

tf.keras.preprocessing.image.save_img("highres.jpg", image)

---

Image Restoration

---

Model: MRNet

Category: Image Restoration

Class: readyml.imagerestoration.MIRNet

Example of use:

from readyml import imagerestoration as rir
import PIL.Image as Image

# Read an image
image_pil = Image.open("../images/mirnet.jpg")

# Instantiate the model class
model = rir.MIRNet()
new_image = model.infer(image_pil)

new_image.save("restored_image.jpg")

---

Text Translation

---

Model: Neural Machine Translation: English to French

Category: Text Translation

Class: readyml.texttranslation.Translation_EnglishToFrench

Example of use:

from readyml import texttranslation as rtt

model = rtt.Translation_EnglishToFrench()
text = model.infer("Hello world!")
print(text)

Result:

Bonjour à tous !

---

Model: Neural Machine Translation: English to German

Category: Text Translation

Class: readyml.texttranslation.Translation_EnglishToGerman

Example of use:

from readyml import texttranslation as rtt

model = rtt.Translation_EnglishToGerman()
text = model.infer("Hello world!")
print(text)

Result:

---

Model: Neural Machine Translation: German to English

Category: Text Translation

Class: readyml.texttranslation.Translation_GermanToEnglish

Example of use:

from readyml import texttranslation as rtt

model = rtt.Translation_GermanToEnglish()
text = model.infer("TBD")
print(text)

Result:

---

Model: Neural Machine Translation: English to Russian

Category: Text Translation

Class: readyml.texttranslation.Translation_EnglishToRussian

Example of use:

from readyml import texttranslation as rtt

model = rtt.Translation_EnglishToRussian()
text = model.infer("Hello world!")
print(text)

Result:

---

Model: Neural Machine Translation: Russian to English

Category: Text Translation

Class: readyml.texttranslation.Translation_RussianToEnglish

Example of use:

from readyml import texttranslation as rtt

model = rtt.Translation_RussianToEnglish()
text = model.infer("")
print(text)

Result: