2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_setup.py:_flush():81] Current SDK version is 0.25.1
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_setup.py:_flush():81] Configure stats pid to 2673297
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_setup.py:_flush():81] Loading settings from environment variables
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_init.py:setup_run_log_directory():717] Logging user logs to /home/sunhc/diffusion/sad/wandb/run-20260417_075829-921e76up/logs/debug.log
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to /home/sunhc/diffusion/sad/wandb/run-20260417_075829-921e76up/logs/debug-internal.log
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_init.py:init():844] calling init triggers
2026-04-17 07:58:29,126 INFO    MainThread:2673297 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
config: {'model': {'type': 'block_ar', 'vocab_size': 50258, 'hidden_size': 768, 'n_blocks': 12, 'n_heads': 12, 'cond_dim': 128, 'max_seq_len': 512, 'block_size': 8, 'dropout': 0.0, 'num_levels': 2, 'level_sizes': [50258, 128]}, 'ancestor': {'lut_path': 'data/ancestor_lut_50257-128_top3_t1.0.pt', 'proto_path': 'data/hierarchy_prototypes_50257-128.pt'}, 'loss': {'lambda_ancestor': 0.0, 'mask_only': False}, 'training': {'seed': 0, 'batch_size': 64, 'num_steps': 10000, 'lr': 0.0003, 'weight_decay': 0.01, 'grad_clip': 1.0, 'dtype': 'bf16', 'curriculum_type': 'linear', 'curriculum_warmup_steps': 5000, 'max_blocks': 64, 'log_interval': 100, 'eval_interval': 5000, 'save_interval': 10000}, 'data': {'dataset': 'openwebtext', 'seq_len': 512, 'cache_dir': 'data/owt_cache', 'num_workers': 4, 'mode': 'subsample'}, 'logging': {'use_wandb': True, 'project': 'sad_block_ar', 'save_dir': 'outputs/sad_block_ar'}, '_wandb': {}}
2026-04-17 07:58:29,127 INFO    MainThread:2673297 [wandb_init.py:init():892] starting backend
2026-04-17 07:58:29,351 INFO    MainThread:2673297 [wandb_init.py:init():895] sending inform_init request
2026-04-17 07:58:29,355 INFO    MainThread:2673297 [wandb_init.py:init():903] backend started and connected
2026-04-17 07:58:29,356 INFO    MainThread:2673297 [wandb_init.py:init():973] updated telemetry
2026-04-17 07:58:29,359 INFO    MainThread:2673297 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout