examples.attn package¶

Submodules¶

examples.attn.mltr30k module¶

Documentation

class examples.attn.mltr30k.AttentionSchedule(n_features: int, warmup_steps: int = 4000)[source]¶

Bases: tensorflow.python.keras.optimizer_v2.learning_rate_schedule.LearningRateSchedule

get_config()[source]¶

class examples.attn.mltr30k.DampedSineSchedule(base_learning_rate: float = 0.0005, min_learning_rate: float = 1e-05, amplitude: float = 50.0, amplitude_decay: float = 80000, angular_frequency: float = 2.2222222222222223e-05, phase: float = - 0.5235987755982988, learning_rate_decay: float = 550000)[source]¶

Bases: tensorflow.python.keras.optimizer_v2.learning_rate_schedule.LearningRateSchedule

get_config()[source]¶

examples.attn.mltr30k.main(args: argparse.Namespace)[source]¶

examples.attn.mltr30k.make_command_line_options()[source]¶

examples.attn.mltr30k.prepare_data(args: argparse.Namespace)[source]¶

examples.attn.mltr30k.setup_model(args: argparse.Namespace)[source]¶

Module contents¶

Documentation