xlm.commands.scaffold_model

Script to scaffold a new external language model for the XLM framework.

Usage: xlm-scaffold [options]

This script creates a complete external model structure with: - Python package with skeleton implementations - Configuration files for all necessary components - Documentation and examples

validate_model_name(name)

Validate and normalize model name.

create_template_context(model_name)

Create template context with all necessary variables.

generate_types_file(model_dir, context)

Generate the types_{model_name}.py file with TypedDict definitions.

generate_model_file(model_dir, context)

Generate the model_{model_name}.py file with the neural network implementation.

generate_loss_file(model_dir, context)

Generate the loss_{model_name}.py file with loss computation.

generate_predictor_file(model_dir, context)

Generate the predictor_{model_name}.py file with inference logic.

generate_datamodule_file(model_dir, context)

Generate the datamodule_.py file with data processing logic.

generate_metrics_file(model_dir, context)

Generate the metrics_{model_name}.py file with metric computation logic.

generate_init_file(model_dir, context, is_core=False)

Generate the init.py file.

generate_config_files(config_dir, context, is_core=False)

Generate all configuration files.

generate_setup_file(model_dir, context)

Generate setup.py for the external model.

generate_documentation(model_dir, context)

Generate README and documentation.

update_xlm_models_file(model_name, xlm_models_path=Path('xlm_models.json'))

Add the new model to the xlm_models.json file.