train_mmlab_segmentation

Train for MMLAB segmentation 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 data 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": "semantic_segmentation",
}) 

# Add train algorithm 
train = wf.add_task(name="train_mmlab_segmentation", 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 'segformer': Name of the model.
• model_config (str) - default 'segformer_mit-b2_8xb2-160k_ade20k-512x512': Name of the config.
• batch_size (int) - default 2: Number of samples processed before the model is updated. Minimum batch_size is 2.
• max_iter (int) - default 1000: Number of training iterations.
• dataset_split_ratio (float) – default '0.9': Divide the dataset into train and evaluation sets ]0, 1[.
• output_folder (str, optional): path to where the model will be saved.
• eval_period (int) - default 100: Number of iterations between 2 evaluations.
• model_weight_file (str, optional): Model weights used as pretrained model. Will use by default mmlab's weights.
• config_file (str, optional): Path to the training config file .yaml. Use it only if you know exactly how mmlab works
• dataset_folder (str, optional): Folder where to save the dataset formatted for mmlab. Is by default in the algorithm directory.

model_name and model_config work by pair. You can print the available possibilities with this code snippet:

from ikomia.dataprocess.workflow import Workflow

# Init your workflow
wf = Workflow()

# Add algorithm
train = wf.add_task(name="train_mmlab_segmentation")

# Get model zoo and print it
model_zoo = train.get_model_zoo()
print(model_zoo)

Note: parameter key and value should be in string format when added to the dictionary.

...
train.set_parameters({
    "param1": "value1",
    "param2": "value2",
})
...