Predict

predict

logger = ColorLog(console, __name__).logger module-attribute

CONFIG_PATH = Path(__file__).parent.parent / 'configs' module-attribute

DiffusionPredictor(config: DictConfig)

Bases: AccelerateDeNovoPredictor

Predictor for the InstaNovo+ model.

refine = config.get('refine', False) instance-attribute

refine_all = config.get('refine_all', True) instance-attribute

refine_threshold = np.log(config.get('refine_threshold', 0.9)) instance-attribute

precursor_tolerance = config.get('filter_precursor_ppm', 50) instance-attribute

load_model() -> Tuple[nn.Module, DictConfig]

Setup the model.

postprocess_dataset(dataset: Dataset) -> Dataset

Load previous predictions for refinement.

setup_data_processor() -> DataProcessor

Setup the data processor.

setup_decoder() -> Decoder

Setup the decoder.

get_predictions(batch: Any) -> dict[str, Any]

Get the predictions for a batch.

CombinedPredictor(config: DictConfig)

Bases: TransformerPredictor

Predictor for the combined InstaNovo+ model.

diffusion_load_model = DiffusionPredictor.load_model class-attribute instance-attribute

diffusion_get_predictions = DiffusionPredictor.get_predictions class-attribute instance-attribute

refine = config.get('refine', False) instance-attribute

refine_all = config.get('refine_all', True) instance-attribute

refine_threshold = np.log(config.get('refine_threshold', 0.9)) instance-attribute

precursor_tolerance = config.get('filter_precursor_ppm', 50) instance-attribute

diffusion_model: nn.Module = self.accelerator.prepare(self.diffusion_model) instance-attribute

load_model() -> Tuple[nn.Module, DictConfig]

Setup the model.

setup_decoder() -> Decoder

Setup the decoder.

get_predictions(batch: Any) -> dict[str, Any]

Get the predictions for a batch.