| { | |
| "best_metric": 0.9777191877365112, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-1_180-samples_config-2_auto/checkpoint-42", | |
| "epoch": 12.0, | |
| "eval_steps": 500, | |
| "global_step": 102, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11764705882352941, | |
| "grad_norm": 1.6483811140060425, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.6016, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.23529411764705882, | |
| "grad_norm": 1.8188143968582153, | |
| "learning_rate": 5e-06, | |
| "loss": 1.5661, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.47058823529411764, | |
| "grad_norm": 1.7118061780929565, | |
| "learning_rate": 1e-05, | |
| "loss": 1.6066, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.7058823529411765, | |
| "grad_norm": 1.8728283643722534, | |
| "learning_rate": 1.5e-05, | |
| "loss": 1.6552, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "grad_norm": 1.846622109413147, | |
| "learning_rate": 2e-05, | |
| "loss": 1.6536, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9411764705882353, | |
| "eval_loss": 1.6148011684417725, | |
| "eval_runtime": 3.175, | |
| "eval_samples_per_second": 11.339, | |
| "eval_steps_per_second": 11.339, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 1.6735312938690186, | |
| "learning_rate": 2.5e-05, | |
| "loss": 1.5927, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.4117647058823528, | |
| "grad_norm": 1.5562609434127808, | |
| "learning_rate": 3e-05, | |
| "loss": 1.4417, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.6470588235294117, | |
| "grad_norm": 1.2983638048171997, | |
| "learning_rate": 3.5e-05, | |
| "loss": 1.353, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.8823529411764706, | |
| "grad_norm": 1.2827444076538086, | |
| "learning_rate": 4e-05, | |
| "loss": 1.2068, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.1806913614273071, | |
| "eval_runtime": 3.2253, | |
| "eval_samples_per_second": 11.162, | |
| "eval_steps_per_second": 11.162, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 2.1176470588235294, | |
| "grad_norm": 0.9886078834533691, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.1324, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 0.9659700989723206, | |
| "learning_rate": 5e-05, | |
| "loss": 1.0622, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.588235294117647, | |
| "grad_norm": 0.8245033621788025, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 1.0729, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.8235294117647056, | |
| "grad_norm": 0.6965786814689636, | |
| "learning_rate": 6e-05, | |
| "loss": 1.099, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "eval_loss": 1.0636965036392212, | |
| "eval_runtime": 3.1493, | |
| "eval_samples_per_second": 11.431, | |
| "eval_steps_per_second": 11.431, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 3.0588235294117645, | |
| "grad_norm": 0.8082783222198486, | |
| "learning_rate": 6.500000000000001e-05, | |
| "loss": 1.0129, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 3.2941176470588234, | |
| "grad_norm": 0.6719391345977783, | |
| "learning_rate": 7e-05, | |
| "loss": 0.96, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 3.5294117647058822, | |
| "grad_norm": 0.6545636653900146, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.9613, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 3.764705882352941, | |
| "grad_norm": 0.6773549914360046, | |
| "learning_rate": 8e-05, | |
| "loss": 0.9415, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.7470717430114746, | |
| "learning_rate": 8.5e-05, | |
| "loss": 0.9298, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.006292700767517, | |
| "eval_runtime": 3.1804, | |
| "eval_samples_per_second": 11.319, | |
| "eval_steps_per_second": 11.319, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 4.235294117647059, | |
| "grad_norm": 0.723663330078125, | |
| "learning_rate": 9e-05, | |
| "loss": 0.8879, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 4.470588235294118, | |
| "grad_norm": 0.7206593751907349, | |
| "learning_rate": 9.5e-05, | |
| "loss": 0.8112, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 4.705882352941177, | |
| "grad_norm": 0.6750977635383606, | |
| "learning_rate": 0.0001, | |
| "loss": 0.8758, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "grad_norm": 0.7167984247207642, | |
| "learning_rate": 9.999238475781957e-05, | |
| "loss": 0.831, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.9411764705882355, | |
| "eval_loss": 0.9777191877365112, | |
| "eval_runtime": 3.162, | |
| "eval_samples_per_second": 11.385, | |
| "eval_steps_per_second": 11.385, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 5.176470588235294, | |
| "grad_norm": 0.6707794070243835, | |
| "learning_rate": 9.99695413509548e-05, | |
| "loss": 0.7714, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 5.411764705882353, | |
| "grad_norm": 0.6780092120170593, | |
| "learning_rate": 9.99314767377287e-05, | |
| "loss": 0.6937, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 5.647058823529412, | |
| "grad_norm": 0.7976924180984497, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.7127, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 5.882352941176471, | |
| "grad_norm": 0.8580948114395142, | |
| "learning_rate": 9.980973490458728e-05, | |
| "loss": 0.7592, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.9820876717567444, | |
| "eval_runtime": 3.2189, | |
| "eval_samples_per_second": 11.184, | |
| "eval_steps_per_second": 11.184, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 6.117647058823529, | |
| "grad_norm": 0.7955440282821655, | |
| "learning_rate": 9.972609476841367e-05, | |
| "loss": 0.7038, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 6.352941176470588, | |
| "grad_norm": 0.8583125472068787, | |
| "learning_rate": 9.962730758206611e-05, | |
| "loss": 0.6277, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 6.588235294117647, | |
| "grad_norm": 0.9593090415000916, | |
| "learning_rate": 9.951340343707852e-05, | |
| "loss": 0.5799, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 6.823529411764706, | |
| "grad_norm": 1.0006083250045776, | |
| "learning_rate": 9.938441702975689e-05, | |
| "loss": 0.5722, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 6.9411764705882355, | |
| "eval_loss": 1.03861665725708, | |
| "eval_runtime": 3.2222, | |
| "eval_samples_per_second": 11.172, | |
| "eval_steps_per_second": 11.172, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 7.0588235294117645, | |
| "grad_norm": 0.9124515652656555, | |
| "learning_rate": 9.924038765061042e-05, | |
| "loss": 0.576, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 7.294117647058823, | |
| "grad_norm": 1.0282957553863525, | |
| "learning_rate": 9.908135917238321e-05, | |
| "loss": 0.5054, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 7.529411764705882, | |
| "grad_norm": 1.3002614974975586, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.4125, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 7.764705882352941, | |
| "grad_norm": 1.231948733329773, | |
| "learning_rate": 9.871850323926177e-05, | |
| "loss": 0.4413, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 1.5320850610733032, | |
| "learning_rate": 9.851478631379982e-05, | |
| "loss": 0.4793, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.2374316453933716, | |
| "eval_runtime": 3.1935, | |
| "eval_samples_per_second": 11.273, | |
| "eval_steps_per_second": 11.273, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 8.235294117647058, | |
| "grad_norm": 1.1022107601165771, | |
| "learning_rate": 9.829629131445342e-05, | |
| "loss": 0.3113, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 8.470588235294118, | |
| "grad_norm": 1.222530722618103, | |
| "learning_rate": 9.806308479691595e-05, | |
| "loss": 0.3486, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 8.705882352941176, | |
| "grad_norm": 1.3544836044311523, | |
| "learning_rate": 9.781523779815179e-05, | |
| "loss": 0.2943, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "grad_norm": 1.7861392498016357, | |
| "learning_rate": 9.755282581475769e-05, | |
| "loss": 0.3157, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 8.941176470588236, | |
| "eval_loss": 1.453012228012085, | |
| "eval_runtime": 3.1504, | |
| "eval_samples_per_second": 11.427, | |
| "eval_steps_per_second": 11.427, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 9.176470588235293, | |
| "grad_norm": 1.2386289834976196, | |
| "learning_rate": 9.727592877996585e-05, | |
| "loss": 0.2149, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 9.411764705882353, | |
| "grad_norm": 1.4310222864151, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.1888, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 9.647058823529411, | |
| "grad_norm": 1.6150720119476318, | |
| "learning_rate": 9.667902132486009e-05, | |
| "loss": 0.1821, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 9.882352941176471, | |
| "grad_norm": 1.6716431379318237, | |
| "learning_rate": 9.635919272833938e-05, | |
| "loss": 0.2078, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.807525873184204, | |
| "eval_runtime": 3.1598, | |
| "eval_samples_per_second": 11.393, | |
| "eval_steps_per_second": 11.393, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 10.117647058823529, | |
| "grad_norm": 1.3494782447814941, | |
| "learning_rate": 9.602524267262203e-05, | |
| "loss": 0.1435, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 10.352941176470589, | |
| "grad_norm": 1.3932836055755615, | |
| "learning_rate": 9.567727288213005e-05, | |
| "loss": 0.106, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 10.588235294117647, | |
| "grad_norm": 1.3322428464889526, | |
| "learning_rate": 9.53153893518325e-05, | |
| "loss": 0.1254, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 10.823529411764707, | |
| "grad_norm": 1.860274076461792, | |
| "learning_rate": 9.493970231495835e-05, | |
| "loss": 0.1328, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 10.941176470588236, | |
| "eval_loss": 2.0376336574554443, | |
| "eval_runtime": 3.1888, | |
| "eval_samples_per_second": 11.29, | |
| "eval_steps_per_second": 11.29, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 11.058823529411764, | |
| "grad_norm": 1.0124211311340332, | |
| "learning_rate": 9.45503262094184e-05, | |
| "loss": 0.1019, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 11.294117647058824, | |
| "grad_norm": 1.3623038530349731, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.089, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 11.529411764705882, | |
| "grad_norm": 1.2576934099197388, | |
| "learning_rate": 9.373098535696979e-05, | |
| "loss": 0.0839, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 11.764705882352942, | |
| "grad_norm": 1.784144401550293, | |
| "learning_rate": 9.330127018922194e-05, | |
| "loss": 0.075, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 1.220141887664795, | |
| "learning_rate": 9.285836503510562e-05, | |
| "loss": 0.0699, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 2.2025742530822754, | |
| "eval_runtime": 3.2125, | |
| "eval_samples_per_second": 11.206, | |
| "eval_steps_per_second": 11.206, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "step": 102, | |
| "total_flos": 7235055271280640.0, | |
| "train_loss": 0.6960301403613651, | |
| "train_runtime": 697.0571, | |
| "train_samples_per_second": 9.755, | |
| "train_steps_per_second": 0.574 | |
| } | |
| ], | |
| "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": 7235055271280640.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |