| { | |
| "best_metric": 1.0136076211929321, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-1_120-samples_config-1_full/checkpoint-66", | |
| "epoch": 13.0, | |
| "eval_steps": 500, | |
| "global_step": 143, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.09090909090909091, | |
| "grad_norm": 1.4600740671157837, | |
| "learning_rate": 1.818181818181818e-06, | |
| "loss": 2.4814, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.4081661701202393, | |
| "learning_rate": 3.636363636363636e-06, | |
| "loss": 2.5532, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 1.4837299585342407, | |
| "learning_rate": 7.272727272727272e-06, | |
| "loss": 2.5951, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.5454545454545454, | |
| "grad_norm": 1.5733487606048584, | |
| "learning_rate": 1.0909090909090909e-05, | |
| "loss": 2.5615, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 1.4437849521636963, | |
| "learning_rate": 1.4545454545454545e-05, | |
| "loss": 2.4794, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "grad_norm": 1.4383858442306519, | |
| "learning_rate": 1.8181818181818182e-05, | |
| "loss": 2.4912, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 2.4122495651245117, | |
| "eval_runtime": 2.2413, | |
| "eval_samples_per_second": 10.708, | |
| "eval_steps_per_second": 10.708, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 1.4773590564727783, | |
| "learning_rate": 2.1818181818181818e-05, | |
| "loss": 2.4687, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 1.2727272727272727, | |
| "grad_norm": 1.3227005004882812, | |
| "learning_rate": 2.5454545454545454e-05, | |
| "loss": 2.4224, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 1.27488112449646, | |
| "learning_rate": 2.909090909090909e-05, | |
| "loss": 2.2447, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 1.6363636363636362, | |
| "grad_norm": 1.2176178693771362, | |
| "learning_rate": 3.272727272727273e-05, | |
| "loss": 2.2032, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 1.176103115081787, | |
| "learning_rate": 3.6363636363636364e-05, | |
| "loss": 2.0095, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.862741231918335, | |
| "learning_rate": 4e-05, | |
| "loss": 1.8242, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.800330638885498, | |
| "eval_runtime": 2.1944, | |
| "eval_samples_per_second": 10.937, | |
| "eval_steps_per_second": 10.937, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 0.9505556225776672, | |
| "learning_rate": 4.3636363636363636e-05, | |
| "loss": 1.7438, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 2.3636363636363638, | |
| "grad_norm": 0.8747909665107727, | |
| "learning_rate": 4.7272727272727275e-05, | |
| "loss": 1.706, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.9589718580245972, | |
| "learning_rate": 5.090909090909091e-05, | |
| "loss": 1.5455, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 2.7272727272727275, | |
| "grad_norm": 1.0078096389770508, | |
| "learning_rate": 5.4545454545454546e-05, | |
| "loss": 1.4537, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.989611804485321, | |
| "learning_rate": 5.818181818181818e-05, | |
| "loss": 1.3712, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.2809544801712036, | |
| "eval_runtime": 2.1892, | |
| "eval_samples_per_second": 10.963, | |
| "eval_steps_per_second": 10.963, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 3.090909090909091, | |
| "grad_norm": 0.7878124713897705, | |
| "learning_rate": 6.181818181818182e-05, | |
| "loss": 1.3423, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.7081627249717712, | |
| "learning_rate": 6.545454545454546e-05, | |
| "loss": 1.186, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 3.4545454545454546, | |
| "grad_norm": 0.7219759821891785, | |
| "learning_rate": 6.90909090909091e-05, | |
| "loss": 1.1499, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.6313851475715637, | |
| "learning_rate": 7.272727272727273e-05, | |
| "loss": 1.148, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 3.8181818181818183, | |
| "grad_norm": 0.6384337544441223, | |
| "learning_rate": 7.636363636363637e-05, | |
| "loss": 1.1229, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.5967474579811096, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0733, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.102345585823059, | |
| "eval_runtime": 2.2143, | |
| "eval_samples_per_second": 10.839, | |
| "eval_steps_per_second": 10.839, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 4.181818181818182, | |
| "grad_norm": 0.5682268142700195, | |
| "learning_rate": 8.363636363636364e-05, | |
| "loss": 1.0328, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.7360040545463562, | |
| "learning_rate": 8.727272727272727e-05, | |
| "loss": 1.0353, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 4.545454545454545, | |
| "grad_norm": 0.6081533432006836, | |
| "learning_rate": 9.090909090909092e-05, | |
| "loss": 1.0128, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.6272777318954468, | |
| "learning_rate": 9.454545454545455e-05, | |
| "loss": 0.9991, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "grad_norm": 0.6040960550308228, | |
| "learning_rate": 9.818181818181818e-05, | |
| "loss": 0.9287, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.031415343284607, | |
| "eval_runtime": 2.2053, | |
| "eval_samples_per_second": 10.883, | |
| "eval_steps_per_second": 10.883, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.6045665144920349, | |
| "learning_rate": 9.999899300364532e-05, | |
| "loss": 0.9217, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 5.2727272727272725, | |
| "grad_norm": 0.6207621693611145, | |
| "learning_rate": 9.99909372761763e-05, | |
| "loss": 0.9, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.6280696392059326, | |
| "learning_rate": 9.997482711915927e-05, | |
| "loss": 0.8373, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 5.636363636363637, | |
| "grad_norm": 0.6763918995857239, | |
| "learning_rate": 9.99506651282272e-05, | |
| "loss": 0.9451, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.7401593923568726, | |
| "learning_rate": 9.991845519630678e-05, | |
| "loss": 0.8251, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 0.7301207184791565, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.8686, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.0136076211929321, | |
| "eval_runtime": 2.2618, | |
| "eval_samples_per_second": 10.611, | |
| "eval_steps_per_second": 10.611, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.6596643924713135, | |
| "learning_rate": 9.982991356370404e-05, | |
| "loss": 0.767, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 6.363636363636363, | |
| "grad_norm": 0.6943769454956055, | |
| "learning_rate": 9.977359612865423e-05, | |
| "loss": 0.7946, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.8311771154403687, | |
| "learning_rate": 9.970925928158274e-05, | |
| "loss": 0.7863, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 6.7272727272727275, | |
| "grad_norm": 0.862726628780365, | |
| "learning_rate": 9.963691338830044e-05, | |
| "loss": 0.7762, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.9442681670188904, | |
| "learning_rate": 9.955657010501806e-05, | |
| "loss": 0.6842, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 1.0364986658096313, | |
| "eval_runtime": 2.239, | |
| "eval_samples_per_second": 10.719, | |
| "eval_steps_per_second": 10.719, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 7.090909090909091, | |
| "grad_norm": 0.8978397250175476, | |
| "learning_rate": 9.946824237646824e-05, | |
| "loss": 0.627, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.9813780188560486, | |
| "learning_rate": 9.937194443381972e-05, | |
| "loss": 0.6837, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 7.454545454545454, | |
| "grad_norm": 0.9994110465049744, | |
| "learning_rate": 9.926769179238466e-05, | |
| "loss": 0.5855, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.9972022771835327, | |
| "learning_rate": 9.915550124911866e-05, | |
| "loss": 0.5865, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 7.818181818181818, | |
| "grad_norm": 1.165213704109192, | |
| "learning_rate": 9.903539087991462e-05, | |
| "loss": 0.6774, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 1.2078843116760254, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.5911, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.0807676315307617, | |
| "eval_runtime": 2.2124, | |
| "eval_samples_per_second": 10.848, | |
| "eval_steps_per_second": 10.848, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 8.181818181818182, | |
| "grad_norm": 1.0711487531661987, | |
| "learning_rate": 9.877148934427037e-05, | |
| "loss": 0.5175, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 1.2479923963546753, | |
| "learning_rate": 9.862774069706346e-05, | |
| "loss": 0.4721, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 8.545454545454545, | |
| "grad_norm": 1.5729472637176514, | |
| "learning_rate": 9.847615725553456e-05, | |
| "loss": 0.442, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 1.233132004737854, | |
| "learning_rate": 9.831676344247342e-05, | |
| "loss": 0.4564, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "grad_norm": 1.4928730726242065, | |
| "learning_rate": 9.814958493905963e-05, | |
| "loss": 0.4827, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 1.2146977186203003, | |
| "eval_runtime": 2.1954, | |
| "eval_samples_per_second": 10.932, | |
| "eval_steps_per_second": 10.932, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 1.2285423278808594, | |
| "learning_rate": 9.797464868072488e-05, | |
| "loss": 0.4189, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 9.272727272727273, | |
| "grad_norm": 1.3862268924713135, | |
| "learning_rate": 9.779198285281325e-05, | |
| "loss": 0.3463, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 1.6900478601455688, | |
| "learning_rate": 9.760161688604008e-05, | |
| "loss": 0.3594, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 9.636363636363637, | |
| "grad_norm": 1.670845627784729, | |
| "learning_rate": 9.740358145174998e-05, | |
| "loss": 0.3092, | |
| "step": 106 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 1.850733995437622, | |
| "learning_rate": 9.719790845697533e-05, | |
| "loss": 0.3153, | |
| "step": 108 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 1.4332870244979858, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.3254, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.4074336290359497, | |
| "eval_runtime": 2.2634, | |
| "eval_samples_per_second": 10.603, | |
| "eval_steps_per_second": 10.603, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 1.3196462392807007, | |
| "learning_rate": 9.676378356149734e-05, | |
| "loss": 0.2496, | |
| "step": 112 | |
| }, | |
| { | |
| "epoch": 10.363636363636363, | |
| "grad_norm": 1.5668597221374512, | |
| "learning_rate": 9.653540160603956e-05, | |
| "loss": 0.2583, | |
| "step": 114 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 1.7147984504699707, | |
| "learning_rate": 9.629952196931901e-05, | |
| "loss": 0.2244, | |
| "step": 116 | |
| }, | |
| { | |
| "epoch": 10.727272727272727, | |
| "grad_norm": 2.276442527770996, | |
| "learning_rate": 9.60561826557425e-05, | |
| "loss": 0.2472, | |
| "step": 118 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 1.6391119956970215, | |
| "learning_rate": 9.580542287160348e-05, | |
| "loss": 0.2263, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_loss": 1.607757568359375, | |
| "eval_runtime": 2.2201, | |
| "eval_samples_per_second": 10.81, | |
| "eval_steps_per_second": 10.81, | |
| "step": 121 | |
| }, | |
| { | |
| "epoch": 11.090909090909092, | |
| "grad_norm": 1.5398436784744263, | |
| "learning_rate": 9.554728301876526e-05, | |
| "loss": 0.1785, | |
| "step": 122 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 1.2709242105484009, | |
| "learning_rate": 9.528180468815155e-05, | |
| "loss": 0.1547, | |
| "step": 124 | |
| }, | |
| { | |
| "epoch": 11.454545454545455, | |
| "grad_norm": 1.4828933477401733, | |
| "learning_rate": 9.50090306530454e-05, | |
| "loss": 0.157, | |
| "step": 126 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 1.658852219581604, | |
| "learning_rate": 9.472900486219769e-05, | |
| "loss": 0.1581, | |
| "step": 128 | |
| }, | |
| { | |
| "epoch": 11.818181818181818, | |
| "grad_norm": 1.8702201843261719, | |
| "learning_rate": 9.444177243274618e-05, | |
| "loss": 0.1609, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 1.8883717060089111, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.1939, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.8584327697753906, | |
| "eval_runtime": 2.2126, | |
| "eval_samples_per_second": 10.847, | |
| "eval_steps_per_second": 10.847, | |
| "step": 132 | |
| }, | |
| { | |
| "epoch": 12.181818181818182, | |
| "grad_norm": 1.0525795221328735, | |
| "learning_rate": 9.384587392471515e-05, | |
| "loss": 0.1172, | |
| "step": 134 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 1.245198369026184, | |
| "learning_rate": 9.353730385598887e-05, | |
| "loss": 0.1173, | |
| "step": 136 | |
| }, | |
| { | |
| "epoch": 12.545454545454545, | |
| "grad_norm": 2.524583339691162, | |
| "learning_rate": 9.322171915289635e-05, | |
| "loss": 0.1475, | |
| "step": 138 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 1.7267895936965942, | |
| "learning_rate": 9.289917066174886e-05, | |
| "loss": 0.1192, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "grad_norm": 1.4045155048370361, | |
| "learning_rate": 9.256971035084785e-05, | |
| "loss": 0.1191, | |
| "step": 142 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_loss": 1.9537454843521118, | |
| "eval_runtime": 2.2484, | |
| "eval_samples_per_second": 10.674, | |
| "eval_steps_per_second": 10.674, | |
| "step": 143 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "step": 143, | |
| "total_flos": 4568364062081024.0, | |
| "train_loss": 0.9211383431212051, | |
| "train_runtime": 503.2208, | |
| "train_samples_per_second": 8.744, | |
| "train_steps_per_second": 1.093 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 550, | |
| "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": 4568364062081024.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |