| { | |
| "best_metric": 1.015079379081726, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-1_120-samples_config-2_full/checkpoint-49", | |
| "epoch": 16.0, | |
| "eval_steps": 500, | |
| "global_step": 88, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.4011845588684082, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 2.5173, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 1.4696787595748901, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 2.595, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 1.342792272567749, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 2.5293, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 2.4960684776306152, | |
| "eval_runtime": 2.2595, | |
| "eval_samples_per_second": 10.622, | |
| "eval_steps_per_second": 10.622, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 1.4615217447280884, | |
| "learning_rate": 2.4e-05, | |
| "loss": 2.5318, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 1.337707281112671, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 2.4877, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 1.3478350639343262, | |
| "learning_rate": 4e-05, | |
| "loss": 2.4062, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.1505632400512695, | |
| "eval_runtime": 2.2783, | |
| "eval_samples_per_second": 10.534, | |
| "eval_steps_per_second": 10.534, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 1.1371965408325195, | |
| "learning_rate": 4.8e-05, | |
| "loss": 2.1431, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 1.0609807968139648, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 2.0298, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.9170504212379456, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.8384, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 1.7334660291671753, | |
| "eval_runtime": 2.2317, | |
| "eval_samples_per_second": 10.754, | |
| "eval_steps_per_second": 10.754, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.8491296768188477, | |
| "learning_rate": 7.2e-05, | |
| "loss": 1.6986, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.9551377296447754, | |
| "learning_rate": 8e-05, | |
| "loss": 1.5698, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.992957353591919, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 1.446, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.3519681692123413, | |
| "eval_runtime": 2.0773, | |
| "eval_samples_per_second": 11.554, | |
| "eval_steps_per_second": 11.554, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.8139051795005798, | |
| "learning_rate": 9.6e-05, | |
| "loss": 1.3135, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.5948655605316162, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 1.2181, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 1.1761022806167603, | |
| "eval_runtime": 2.2085, | |
| "eval_samples_per_second": 10.867, | |
| "eval_steps_per_second": 10.867, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.5459035634994507, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 1.1256, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.46117210388183594, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 1.0769, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.4251512587070465, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 1.0795, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.0929523706436157, | |
| "eval_runtime": 2.059, | |
| "eval_samples_per_second": 11.656, | |
| "eval_steps_per_second": 11.656, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.4119170010089874, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 1.0286, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.46514835953712463, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 1.0456, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.4424739181995392, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.9652, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 1.0544281005859375, | |
| "eval_runtime": 2.0674, | |
| "eval_samples_per_second": 11.609, | |
| "eval_steps_per_second": 11.609, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.4187925159931183, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.946, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.444961279630661, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.8848, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.4570990800857544, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.9317, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.0245299339294434, | |
| "eval_runtime": 2.0697, | |
| "eval_samples_per_second": 11.596, | |
| "eval_steps_per_second": 11.596, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.4696873426437378, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.8783, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.42547792196273804, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.8531, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 1.015079379081726, | |
| "eval_runtime": 2.0659, | |
| "eval_samples_per_second": 11.617, | |
| "eval_steps_per_second": 11.617, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.4868246614933014, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.8371, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.45950591564178467, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.8066, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.5040105581283569, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.7657, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.0206605195999146, | |
| "eval_runtime": 2.0835, | |
| "eval_samples_per_second": 11.519, | |
| "eval_steps_per_second": 11.519, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.49457821249961853, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.7622, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.5479050278663635, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.7314, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.5930130481719971, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.7049, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 1.0355324745178223, | |
| "eval_runtime": 2.0639, | |
| "eval_samples_per_second": 11.628, | |
| "eval_steps_per_second": 11.628, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.6274011135101318, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.6268, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.6222364902496338, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.6432, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.7149001955986023, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.6405, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.0796736478805542, | |
| "eval_runtime": 2.0635, | |
| "eval_samples_per_second": 11.631, | |
| "eval_steps_per_second": 11.631, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.6841269135475159, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.5615, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.7190103530883789, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.5447, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 1.1631149053573608, | |
| "eval_runtime": 2.0685, | |
| "eval_samples_per_second": 11.603, | |
| "eval_steps_per_second": 11.603, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.8411610126495361, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.5143, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.805773913860321, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.4693, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.9177535176277161, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.4351, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.285775065422058, | |
| "eval_runtime": 2.0598, | |
| "eval_samples_per_second": 11.652, | |
| "eval_steps_per_second": 11.652, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.8271195292472839, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.4023, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 1.0611273050308228, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.359, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 1.004104733467102, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.3579, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 1.4291505813598633, | |
| "eval_runtime": 2.1082, | |
| "eval_samples_per_second": 11.384, | |
| "eval_steps_per_second": 11.384, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 15.272727272727273, | |
| "grad_norm": 0.9628243446350098, | |
| "learning_rate": 8.397206521307584e-05, | |
| "loss": 0.2763, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 15.636363636363637, | |
| "grad_norm": 1.1773602962493896, | |
| "learning_rate": 8.293447300593402e-05, | |
| "loss": 0.2725, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 1.3530744314193726, | |
| "learning_rate": 8.18711994874345e-05, | |
| "loss": 0.2641, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.6355339288711548, | |
| "eval_runtime": 2.12, | |
| "eval_samples_per_second": 11.321, | |
| "eval_steps_per_second": 11.321, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 88, | |
| "total_flos": 5622601906913280.0, | |
| "train_loss": 1.0808867636052044, | |
| "train_runtime": 602.3235, | |
| "train_samples_per_second": 7.305, | |
| "train_steps_per_second": 0.415 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 250, | |
| "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": 5622601906913280.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |