Skip to content

Training and End-to-End dataset (Fridge Objects)

In this example, we are training the Fridge Objects dataset using either Fastai or Pytorch-Lightning training loop

# pip install icevision[all] icedata

from icevision.all import *

url = "https://cvbp-secondary.z19.web.core.windows.net/datasets/object_detection/odFridgeObjects.zip"
dest_dir = "fridge"

# Loading Data
data_dir = icedata.load_data(url, dest_dir)

# Parser
class_map = ClassMap(["milk_bottle", "carton", "can", "water_bottle"])
parser = parsers.voc(annotations_dir=data_dir / "odFridgeObjects/annotations/",
                    images_dir=data_dir / "odFridgeObjects/images",
                    class_map=class_map)

# Records
train_records, valid_records = parser.parse()

# Transforms
train_tfms = tfms.A.Adapter([*tfms.A.aug_tfms(size=384, presize=512), tfms.A.Normalize()])
valid_tfms = tfms.A.Adapter([*tfms.A.resize_and_pad(384), tfms.A.Normalize()])

# Datasets
train_ds = Dataset(train_records, train_tfms)
valid_ds = Dataset(valid_records, valid_tfms)

# DataLoaders
train_dl = efficientdet.train_dl(train_ds, batch_size=16, num_workers=4, shuffle=True)
valid_dl = efficientdet.valid_dl(valid_ds, batch_size=16, num_workers=4, shuffle=False)

# Model and Metrics
model = efficientdet.model(model_name="tf_efficientdet_lite0", num_classes=len(class_map), img_size=size)
metrics = [COCOMetric(metric_type=COCOMetricType.bbox)]

# Training using Fastai
learn = efficientdet.fastai.learner(dls=[train_dl, valid_dl], model=model, metrics=metrics)
learn.fine_tune(50, 1e-2, freeze_epochs=20)

# Inference
# DataLoader
infer_dl = efficientdet.infer_dl(valid_ds, batch_size=8)
# Predict
samples, preds = efficientdet.predict_dl(model, infer_dl)
# Show samples
imgs = [sample["img"] for sample in samples]
show_preds(
    imgs=imgs[:6],
    preds=preds[:6],
    class_map=class_map,
    denormalize_fn=denormalize_imagenet,
    ncols=3,
)

# pip install icevision[all] icedata

from icevision.all import *
import icedata

url = "https://cvbp-secondary.z19.web.core.windows.net/datasets/object_detection/odFridgeObjects.zip"
dest_dir = "fridge"

# Loading Data
data_dir = icedata.load_data(url, dest_dir)

# Parser
class_map = ClassMap(["milk_bottle", "carton", "can", "water_bottle"])
parser = parsers.voc(annotations_dir=data_dir / "odFridgeObjects/images/",
                    images_dir=data_dir / "odFridgeObjects/annotations",
                    class_map=class_map)

# Records
train_records, valid_records = parser.parse()

train_tfms = tfms.A.Adapter([*tfms.A.aug_tfms(size=384, presize=512), tfms.A.Normalize()])
valid_tfms = tfms.A.Adapter([*tfms.A.resize_and_pad(384), tfms.A.Normalize()])

# Datasets
train_ds = Dataset(train_records, train_tfms)
valid_ds = Dataset(valid_records, valid_tfms)

# DataLoaders
train_dl = efficientdet.train_dl(train_ds, batch_size=16, num_workers=4, shuffle=True)
valid_dl = efficientdet.valid_dl(valid_ds, batch_size=16, num_workers=4, shuffle=False)

# Model and Metrics
model = efficientdet.model(model_name="tf_efficientdet_lite0", num_classes=len(class_map), img_size=size)
metrics = [COCOMetric(metric_type=COCOMetricType.bbox)]

# Training using Pytorch Lightning
class LightModel(efficientdet.lightning.ModelAdapter):
    def configure_optimizers(self):
        return SGD(self.parameters(), lr=1e-2)    

light_model = LightModel(model, metrics=metrics)
trainer = pl.Trainer(max_epochs=70, gpus=1)
trainer.fit(light_model, train_dl, valid_dl)

# Inference
# DataLoader
infer_dl = efficientdet.infer_dl(valid_ds, batch_size=8)
# Predict
samples, preds = efficientdet.predict_dl(model, infer_dl)
# Show samples
imgs = [sample["img"] for sample in samples]
show_preds(
    imgs=imgs[:6],
    preds=preds[:6],
    class_map=class_map,
    denormalize_fn=denormalize_imagenet,
    ncols=3,
)