data: mode: std features_dir: ./data/features strains_path: ./data/strains.parquet metadata_path: ./data/beatmaps.parquet dataset_seed: 42 sample_size: 700000 max_seq_len: 4096 val_split: 0 min_sr: 2.0 max_sr: 14.0 dataloader: num_workers: 2 pin_memory: true persistent_workers: true prefetch_factor: 2 model: d_model: 512 n_heads: 8 n_layers: 9 dim_feedforward: 1408 dropout: 0.1 local_attention_window: 128 local_attention_block_size: 256 global_attention_layers: - 2 - 5 - 8 feature_token_dim: 16 runtime: compile_model: true compile_mode: default activation_checkpointing: false training: seed: 42 trainer: epochs: 20 batch_size: 64 max_tokens: 32768 grad_accum: 2 grad_clip: 1.0 precision: bf16-mixed save_checkpoints: true optimizer: adamw_lr: 0.0002 adamw_wd: 0.05 adamw_betas: - 0.9 - 0.95 muon_lr: 0.01 muon_wd: 0.01 scheduler: warmup_ratio: 0.01 min_lr_ratio: 0.005 section_contrast_weight: 0.02 masking: ratio: 0.3 mean_span_length: 2