{ "next_trial_id": 11, "trials": { "0": { "trial_id": 0, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 4000, "val_games": 500, "min_ply": 10, "total_steps": 100, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.05, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": null, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "float16", "no_compile": false, "sdpa_math": true, "num_workers": 0, "device": "cuda", "log_dir": "/workspace/logs/trial_0000", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": null, "bottleneck_dim": 16, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0000/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 4000, "val_games": 500, "min_ply": 10, "total_steps": 100, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.05, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": null, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "float16", "no_compile": false, "sdpa_math": true, "num_workers": 0, "device": "cuda", "log_dir": "/workspace/logs/trial_0000", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": null, "bottleneck_dim": 16, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "completed", "pid": 60913, "gpu_id": 0, "start_time": 1775346146.9661582, "end_time": 1775346192.2253375, "current_step": 100, "total_steps": 100, "steps_per_sec": 2.499625056241564, "last_train_loss": 2.6580834102494304, "last_train_acc": 0.2578537154243756, "best_val_loss": 2.6815053782271594, "best_accuracy": 0.25373134694665184, "actual_param_count": 262144, "log_path": "/workspace/sweep_results/trial_0000.log", "run_dir": "/workspace/logs/trial_0000/bottleneck_20260404_234228_keen-bison", "optuna_number": null, "notes": "MCP smoke test: 256k bottleneck, 100 steps, MATH+compile+fp16", "tags": [ "mcp-test", "bottleneck-small" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "17h18m" }, "1": { "trial_id": 1, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 4000, "val_games": 500, "min_ply": 10, "total_steps": 30, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.05, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": null, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "float16", "no_compile": false, "sdpa_math": true, "num_workers": 0, "device": "cuda", "log_dir": "/workspace/logs/trial_0000", "hf_repo": null, "local_checkpoints": true, "resume": "/workspace/logs/trial_0000/bottleneck_20260404_234228_keen-bison/checkpoints/best", "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": null, "bottleneck_dim": 16, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0001/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 4000, "val_games": 500, "min_ply": 10, "total_steps": 30, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.05, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": null, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "float16", "no_compile": false, "sdpa_math": true, "num_workers": 0, "device": "cuda", "log_dir": "/workspace/logs/trial_0000", "hf_repo": null, "local_checkpoints": true, "resume": "/workspace/logs/trial_0000/bottleneck_20260404_234228_keen-bison/checkpoints/best", "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": null, "bottleneck_dim": 16, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 61357, "gpu_id": 0, "start_time": 1775346206.78868, "end_time": 1775346209.3660913, "current_step": 0, "total_steps": 30, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0001.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "mcp-test", "bottleneck-small" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "17h17m" }, "2": { "trial_id": 2, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 3500000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 20000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": false, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0002/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 3500000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 20000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": false, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 71057, "gpu_id": 0, "start_time": 1775351012.9564307, "end_time": 1775351080.0435607, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0002.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "scaling", "20M" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h57m" }, "3": { "trial_id": 3, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 20000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": false, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0003/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 20000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": false, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 72205, "gpu_id": 0, "start_time": 1775351111.2472177, "end_time": 1775351835.388752, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0003.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "scaling", "20M", "full-data" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h55m" }, "4": { "trial_id": 4, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0004/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 4, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 73543, "gpu_id": 0, "start_time": 1775351848.6815178, "end_time": 1775352133.8517451, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0004.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "scaling", "20M", "eager" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h43m" }, "5": { "trial_id": 5, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0005/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 50, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 74605, "gpu_id": 0, "start_time": 1775352154.8011737, "end_time": 1775352288.1328025, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0005.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "scaling", "20M", "workers2" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h38m" }, "6": { "trial_id": 6, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 100, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0006/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 100, "batch_size": 64, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "completed", "pid": 75207, "gpu_id": 0, "start_time": 1775352312.4274437, "end_time": 1775352342.5313256, "current_step": 0, "total_steps": 100, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0006.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "diagnostic", "small-bs" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h35m" }, "7": { "trial_id": 7, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 100, "batch_size": 128, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0007/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 100, "batch_size": 128, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 20, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "completed", "pid": 75775, "gpu_id": 0, "start_time": 1775352384.985535, "end_time": 1775352420.1244109, "current_step": 0, "total_steps": 100, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0007.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "diagnostic", "bs128" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h34m" }, "8": { "trial_id": 8, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 50, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 10, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0008/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 5000, "min_ply": 10, "total_steps": 50, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 10, "pause_after_steps": null, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "completed", "pid": 76342, "gpu_id": 0, "start_time": 1775352440.4786208, "end_time": 1775352475.6319711, "current_step": 0, "total_steps": 50, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0008.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "diagnostic", "bs256" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h33m" }, "9": { "trial_id": 9, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 10, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0009/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 10, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "killed", "pid": 77221, "gpu_id": 0, "start_time": 1775352573.273321, "end_time": 1775352817.925155, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0009.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "scaling", "20M", "bs256" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h31m" }, "10": { "trial_id": 10, "strategy": "bottleneck", "params": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 100, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "cli_command": [ "python3", "/opt/pawn/scripts/train.py", "--config", "/workspace/logs/trial_0010/run_config.json" ], "config": { "elo_min": 1800, "elo_max": 1900, "max_games": 30000000, "val_games": 50000, "min_ply": 10, "total_steps": 200000, "batch_size": 256, "lr": 0.0003, "weight_decay": 0.0, "warmup_frac": 0.02, "warmup_steps": null, "max_grad_norm": 1.0, "patience": 9999, "eval_interval": 2500, "log_interval": 100, "pause_after_steps": 15000, "mate_boost": 0.0, "no_outcome_token": false, "discard_ply_limit": false, "amp_dtype": "bfloat16", "no_compile": true, "sdpa_math": true, "num_workers": 2, "device": "cuda", "log_dir": "/workspace/logs", "hf_repo": null, "local_checkpoints": true, "resume": null, "wandb": false, "cache_dir": "/dev/shm/pawn_cache", "run_type": "adapter", "strategy": "bottleneck", "checkpoint": "thomas-schweich/pawn-base", "pgn": "thomas-schweich/pawn-lichess-full", "adapter_layers": "4,5,6,7", "bottleneck_dim": 2440, "no_adapt_attn": false, "no_adapt_ffn": false, "lora_rank": null, "lora_targets": null, "lora_ffn": false, "density": null, "sparse_targets": null, "sparse_ffn": false, "use_output_film": false, "rosa_mode": null, "rosa_warmup_steps": 128, "mask_samples": 32, "grad_alpha": 2, "unfreeze_layers": null, "d_model": null, "n_layers": null, "n_heads": null, "epochs": 9999, "val_every": 9999 }, "status": "completed", "pid": 78502, "gpu_id": 0, "start_time": 1775352861.543256, "end_time": 1775356243.7400186, "current_step": 0, "total_steps": 200000, "steps_per_sec": 0.0, "last_train_loss": null, "last_train_acc": null, "best_val_loss": null, "best_accuracy": null, "actual_param_count": null, "log_path": "/workspace/sweep_results/trial_0010.log", "run_dir": null, "optuna_number": null, "notes": "", "tags": [ "18h-push", "20M", "bottleneck", "scaling" ], "eta_seconds": null, "eta_human": "?", "elapsed_human": "15h26m" } }, "event_seq": 27, "start_time": 1775342775.318943, "cost_per_hour": 2.49 }