| { | |
| "best_metric": 0.872577428817749, | |
| "best_model_checkpoint": "data/Llama-31-8B_task-1_180-samples_config-2_full/checkpoint-76", | |
| "epoch": 16.0, | |
| "eval_steps": 500, | |
| "global_step": 136, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11764705882352941, | |
| "grad_norm": 1.8353525400161743, | |
| "learning_rate": 2.5e-06, | |
| "loss": 2.4537, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 1.8774621486663818, | |
| "learning_rate": 5e-06, | |
| "loss": 2.3789, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 1.7344152927398682, | |
| "learning_rate": 1e-05, | |
| "loss": 2.4305, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 1.6874200105667114, | |
| "learning_rate": 1.5e-05, | |
| "loss": 2.4333, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 1.6447762250900269, | |
| "learning_rate": 2e-05, | |
| "loss": 2.375, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 2.287229537963867, | |
| "eval_runtime": 14.4998, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 1.3521522283554077, | |
| "learning_rate": 2.5e-05, | |
| "loss": 2.2456, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 1.16240656375885, | |
| "learning_rate": 3e-05, | |
| "loss": 2.193, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 1.037164330482483, | |
| "learning_rate": 3.5e-05, | |
| "loss": 2.1086, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 1.6905419826507568, | |
| "learning_rate": 4e-05, | |
| "loss": 1.9761, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.8512320518493652, | |
| "eval_runtime": 14.4943, | |
| "eval_samples_per_second": 2.484, | |
| "eval_steps_per_second": 2.484, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.738078236579895, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.8818, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.7864554524421692, | |
| "learning_rate": 5e-05, | |
| "loss": 1.784, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.913241982460022, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.6715, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.9287745952606201, | |
| "learning_rate": 6e-05, | |
| "loss": 1.5652, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.3758809566497803, | |
| "eval_runtime": 14.4994, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 1.0956100225448608, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.3646, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.7750384211540222, | |
| "learning_rate": 7e-05, | |
| "loss": 1.2983, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.5585227012634277, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 1.1593, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.4572974145412445, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0819, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.4609680473804474, | |
| "learning_rate": 8.5e-05, | |
| "loss": 1.0274, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.0473541021347046, | |
| "eval_runtime": 14.5403, | |
| "eval_samples_per_second": 2.476, | |
| "eval_steps_per_second": 2.476, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.5395373702049255, | |
| "learning_rate": 9e-05, | |
| "loss": 1.0132, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.3620060384273529, | |
| "learning_rate": 9.5e-05, | |
| "loss": 1.011, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.3471994400024414, | |
| "learning_rate": 0.0001, | |
| "loss": 0.996, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.3922196924686432, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.9811, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 0.987396776676178, | |
| "eval_runtime": 14.4999, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.3318175971508026, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.9489, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.3675948679447174, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.9413, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.43041494488716125, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.8816, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.4317922592163086, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.9464, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.9454231262207031, | |
| "eval_runtime": 14.4987, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.4054967164993286, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.8859, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.4406815767288208, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.8887, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.5343546271324158, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.8876, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 0.6281182765960693, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.8491, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 0.9103401899337769, | |
| "eval_runtime": 14.5211, | |
| "eval_samples_per_second": 2.479, | |
| "eval_steps_per_second": 2.479, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.6439505219459534, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.86, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 0.6379028558731079, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.8241, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 0.5003238916397095, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.7798, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 0.39182040095329285, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.7739, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.3933279812335968, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.8043, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.8865704536437988, | |
| "eval_runtime": 14.5262, | |
| "eval_samples_per_second": 2.478, | |
| "eval_steps_per_second": 2.478, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 0.38726237416267395, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.7605, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 0.43751487135887146, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.7692, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 0.46520718932151794, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.6911, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 0.5554521679878235, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.7279, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 0.872577428817749, | |
| "eval_runtime": 14.5135, | |
| "eval_samples_per_second": 2.48, | |
| "eval_steps_per_second": 2.48, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 0.4435398578643799, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.6799, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 0.4906449615955353, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.6431, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 0.5347185134887695, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.6711, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 0.5857811570167542, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.6704, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.8894259929656982, | |
| "eval_runtime": 14.4989, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 0.5464484095573425, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.6006, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 0.614228367805481, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.5622, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 0.6241901516914368, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.604, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 0.6841773390769958, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.6154, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 0.9156128764152527, | |
| "eval_runtime": 14.4889, | |
| "eval_samples_per_second": 2.485, | |
| "eval_steps_per_second": 2.485, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 0.6057692170143127, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.5998, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 0.7102330923080444, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.519, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 0.7177324295043945, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.5065, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 0.8749062418937683, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.4775, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.8146276473999023, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.4469, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.9676671624183655, | |
| "eval_runtime": 14.4879, | |
| "eval_samples_per_second": 2.485, | |
| "eval_steps_per_second": 2.485, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.235294117647058, | |
| "grad_norm": 0.8840669989585876, | |
| "learning_rate": 9.24024048078213e-05, | |
| "loss": 0.4391, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 12.470588235294118, | |
| "grad_norm": 1.4158061742782593, | |
| "learning_rate": 9.193352839727121e-05, | |
| "loss": 0.4116, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 12.705882352941176, | |
| "grad_norm": 1.0295145511627197, | |
| "learning_rate": 9.145187862775209e-05, | |
| "loss": 0.3659, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "grad_norm": 1.1074610948562622, | |
| "learning_rate": 9.09576022144496e-05, | |
| "loss": 0.3873, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 12.941176470588236, | |
| "eval_loss": 1.0959347486495972, | |
| "eval_runtime": 14.4945, | |
| "eval_samples_per_second": 2.484, | |
| "eval_steps_per_second": 2.484, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 13.176470588235293, | |
| "grad_norm": 1.0288474559783936, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.3203, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 13.411764705882353, | |
| "grad_norm": 1.4014495611190796, | |
| "learning_rate": 8.993177550236464e-05, | |
| "loss": 0.2736, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 13.647058823529411, | |
| "grad_norm": 1.1403602361679077, | |
| "learning_rate": 8.940053768033609e-05, | |
| "loss": 0.2805, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 13.882352941176471, | |
| "grad_norm": 1.116623044013977, | |
| "learning_rate": 8.885729807284856e-05, | |
| "loss": 0.3086, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.235299825668335, | |
| "eval_runtime": 14.4876, | |
| "eval_samples_per_second": 2.485, | |
| "eval_steps_per_second": 2.485, | |
| "step": 119 | |
| }, | |
| { | |
| "epoch": 14.117647058823529, | |
| "grad_norm": 1.0906065702438354, | |
| "learning_rate": 8.83022221559489e-05, | |
| "loss": 0.2841, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 14.352941176470589, | |
| "grad_norm": 1.6509333848953247, | |
| "learning_rate": 8.773547901113862e-05, | |
| "loss": 0.2144, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 14.588235294117647, | |
| "grad_norm": 1.1187314987182617, | |
| "learning_rate": 8.715724127386972e-05, | |
| "loss": 0.2273, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 14.823529411764707, | |
| "grad_norm": 1.3310633897781372, | |
| "learning_rate": 8.656768508095853e-05, | |
| "loss": 0.1952, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 14.941176470588236, | |
| "eval_loss": 1.3687325716018677, | |
| "eval_runtime": 14.4971, | |
| "eval_samples_per_second": 2.483, | |
| "eval_steps_per_second": 2.483, | |
| "step": 127 | |
| }, | |
| { | |
| "epoch": 15.058823529411764, | |
| "grad_norm": 1.2637333869934082, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.1943, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 15.294117647058824, | |
| "grad_norm": 1.0221024751663208, | |
| "learning_rate": 8.535533905932738e-05, | |
| "loss": 0.1525, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 15.529411764705882, | |
| "grad_norm": 1.5778526067733765, | |
| "learning_rate": 8.473291852294987e-05, | |
| "loss": 0.1474, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 15.764705882352942, | |
| "grad_norm": 1.036356806755066, | |
| "learning_rate": 8.409991800312493e-05, | |
| "loss": 0.1495, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 1.134097933769226, | |
| "learning_rate": 8.345653031794292e-05, | |
| "loss": 0.1665, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.4522809982299805, | |
| "eval_runtime": 14.4954, | |
| "eval_samples_per_second": 2.484, | |
| "eval_steps_per_second": 2.484, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 136, | |
| "total_flos": 3.4636862551228416e+16, | |
| "train_loss": 0.9021811117144192, | |
| "train_runtime": 2564.085, | |
| "train_samples_per_second": 2.652, | |
| "train_steps_per_second": 0.156 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 400, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.4636862551228416e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |