train_rf_detr

Train RF-DETR object detection models.

🚀 Use with Ikomia API

1. Install Ikomia API

We strongly recommend using a virtual environment. If you're not sure where to start, we offer a tutorial here.

pip install ikomia

2. Create your workflow

from ikomia.dataprocess.workflow import Workflow

# Init your workflow
wf = Workflow()    

# Add dataset loader
coco = wf.add_task(name="dataset_coco")

coco.set_parameters({
    "json_file": "path/to/json/annotation/file",
    "image_folder": "path/to/image/folder",
    "task": "detection",
}) 

# Add training algorithm
train = wf.add_task(name="train_rf_detr", auto_connect=True)

# Launch your training on your data
wf.run()

☀️ Use with Ikomia Studio

Ikomia Studio offers a friendly UI with the same features as the API.

• If you haven't started using Ikomia Studio yet, download and install it from this page.
• For additional guidance on getting started with Ikomia Studio, check out this blog post.

📝 Set algorithm parameters

• model_name (str) - default 'rf-detr-base': Name of the RF-DETR pre-trained model. Other model available:
  • rf-detr-large
• batch_size (int) - default '8': Number of samples processed before the model is updated.
• epochs (int) - default '100': Number of complete passes through the training dataset.
• dataset_split_ratio (float) – default '0.9': Divide the dataset into train and evaluation sets ]0, 1[.
• input_size (int) - default '560': Size of the input image.
• weight_decay (float) - default '0.000125': Amount of weight decay, regularization method.
• workers (int) - default '0': Number of worker threads for data loading (per RANK if DDP).
• lr (float) - default '0.00025': Initial learning rate. Adjusting this value is crucial for the optimization process, influencing how rapidly model weights are updated.
• lr_encoder (float) - default '1.5e-4': Separate learning rate for the encoder parameters. Allows fine-tuning at a different rate than the rest of the model.
• output_folder (str, optional): path to where the model will be saved.
• early_stopping (bool) - default 'False': Whether to enable early stopping during training. This stops training if performance stops improving after a certain number of epochs.
• early_stopping_patience(int) - default '10': Number of consecutive validation checks with no improvement before early stopping is triggered. Only applicable if early_stopping=True.

Parameters should be in strings format when added to the dictionary.

from ikomia.dataprocess.workflow import Workflow

# Init your workflow
wf = Workflow()    

# Add dataset loader
coco = wf.add_task(name="dataset_coco")

coco.set_parameters({
    "json_file": "path/to/json/annotation/file",
    "image_folder": "path/to/image/folder",
    "task": "detection",
}) 

# Add training algorithm
train = wf.add_task(name="train_rf_detr", auto_connect=True)

train.set_parameters({
    "model_name": "dfine_m",
    "epochs": "100",
    "batch_size": "6",
    "input_size": "560",
    "dataset_split_ratio": "0.9",
    "workers": "0",  # Recommended to set to 0 if you are using Windows
    "weight_decay": "1e-4",
    "lr": " 1e-4",
    "output_folder": "Path/To/Output/Folder", # Default folder : runs     
})

# Launch your training on your data
wf.run()