| { | |
| "best_metric": 0.8693992495536804, | |
| "best_model_checkpoint": "data/Llama-31-8B_task-1_120-samples_config-2_full/checkpoint-66", | |
| "epoch": 18.90909090909091, | |
| "eval_steps": 500, | |
| "global_step": 104, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.942243218421936, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 2.4963, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 1.9989137649536133, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 2.515, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 1.5480235815048218, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 2.448, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 2.3731963634490967, | |
| "eval_runtime": 9.6256, | |
| "eval_samples_per_second": 2.493, | |
| "eval_steps_per_second": 2.493, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 1.586003065109253, | |
| "learning_rate": 2.4e-05, | |
| "loss": 2.4028, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 1.5806779861450195, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 2.3232, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 1.2769224643707275, | |
| "learning_rate": 4e-05, | |
| "loss": 2.2361, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 2.027904510498047, | |
| "eval_runtime": 9.6154, | |
| "eval_samples_per_second": 2.496, | |
| "eval_steps_per_second": 2.496, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 1.6216256618499756, | |
| "learning_rate": 4.8e-05, | |
| "loss": 2.0218, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.9748833775520325, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.9447, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.8017610311508179, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.8253, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 1.7251297235488892, | |
| "eval_runtime": 9.6206, | |
| "eval_samples_per_second": 2.495, | |
| "eval_steps_per_second": 2.495, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.9146191477775574, | |
| "learning_rate": 7.2e-05, | |
| "loss": 1.6844, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 1.028951644897461, | |
| "learning_rate": 8e-05, | |
| "loss": 1.5093, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 1.1190450191497803, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 1.3791, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.2330317497253418, | |
| "eval_runtime": 9.6142, | |
| "eval_samples_per_second": 2.496, | |
| "eval_steps_per_second": 2.496, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.6760878562927246, | |
| "learning_rate": 9.6e-05, | |
| "loss": 1.1944, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.4904247522354126, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 1.0878, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 1.0336649417877197, | |
| "eval_runtime": 9.6445, | |
| "eval_samples_per_second": 2.488, | |
| "eval_steps_per_second": 2.488, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.47956767678260803, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 0.9978, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.37655767798423767, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.9748, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.3446647524833679, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.9771, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.973893404006958, | |
| "eval_runtime": 9.6397, | |
| "eval_samples_per_second": 2.49, | |
| "eval_steps_per_second": 2.49, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.32855767011642456, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.9427, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.37166285514831543, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.9719, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.35821229219436646, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.8967, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.9425709843635559, | |
| "eval_runtime": 9.7372, | |
| "eval_samples_per_second": 2.465, | |
| "eval_steps_per_second": 2.465, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.41391488909721375, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.8814, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.4022800326347351, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.8303, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.40724870562553406, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.8815, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.9129160046577454, | |
| "eval_runtime": 9.7169, | |
| "eval_samples_per_second": 2.47, | |
| "eval_steps_per_second": 2.47, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.45184680819511414, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.8283, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.5330278277397156, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.816, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.8952023386955261, | |
| "eval_runtime": 9.6573, | |
| "eval_samples_per_second": 2.485, | |
| "eval_steps_per_second": 2.485, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.5944344997406006, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.8059, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.6521593332290649, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.7773, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.6768603920936584, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.748, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.8761978149414062, | |
| "eval_runtime": 9.6144, | |
| "eval_samples_per_second": 2.496, | |
| "eval_steps_per_second": 2.496, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.6031374335289001, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.7397, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.47089871764183044, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.7211, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.44608962535858154, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.6939, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.8726906776428223, | |
| "eval_runtime": 9.6232, | |
| "eval_samples_per_second": 2.494, | |
| "eval_steps_per_second": 2.494, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.47155243158340454, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.6328, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.5070695877075195, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.6563, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.5034713745117188, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.6449, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.8693992495536804, | |
| "eval_runtime": 9.6261, | |
| "eval_samples_per_second": 2.493, | |
| "eval_steps_per_second": 2.493, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.513129711151123, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.5736, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.5481001138687134, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.5874, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 0.8921452164649963, | |
| "eval_runtime": 9.6307, | |
| "eval_samples_per_second": 2.492, | |
| "eval_steps_per_second": 2.492, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.5650292634963989, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.5501, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.6034680008888245, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.518, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.7035646438598633, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.4934, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.9428841471672058, | |
| "eval_runtime": 9.6183, | |
| "eval_samples_per_second": 2.495, | |
| "eval_steps_per_second": 2.495, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.6939181089401245, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.4695, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 0.8460732102394104, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.4428, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 0.8051614165306091, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.4382, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 1.008301854133606, | |
| "eval_runtime": 9.6147, | |
| "eval_samples_per_second": 2.496, | |
| "eval_steps_per_second": 2.496, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 15.272727272727273, | |
| "grad_norm": 0.6874304413795471, | |
| "learning_rate": 8.397206521307584e-05, | |
| "loss": 0.3561, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 15.636363636363637, | |
| "grad_norm": 1.0137032270431519, | |
| "learning_rate": 8.293447300593402e-05, | |
| "loss": 0.3449, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.9092028737068176, | |
| "learning_rate": 8.18711994874345e-05, | |
| "loss": 0.347, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.0591822862625122, | |
| "eval_runtime": 9.6202, | |
| "eval_samples_per_second": 2.495, | |
| "eval_steps_per_second": 2.495, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.363636363636363, | |
| "grad_norm": 0.9383953213691711, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.2896, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 16.727272727272727, | |
| "grad_norm": 1.1991755962371826, | |
| "learning_rate": 7.967094433018508e-05, | |
| "loss": 0.2565, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 16.90909090909091, | |
| "eval_loss": 1.145786166191101, | |
| "eval_runtime": 9.6302, | |
| "eval_samples_per_second": 2.492, | |
| "eval_steps_per_second": 2.492, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 17.09090909090909, | |
| "grad_norm": 1.0061393976211548, | |
| "learning_rate": 7.85356783842216e-05, | |
| "loss": 0.2364, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 17.454545454545453, | |
| "grad_norm": 1.079446792602539, | |
| "learning_rate": 7.737816117462752e-05, | |
| "loss": 0.1962, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 17.818181818181817, | |
| "grad_norm": 1.1677623987197876, | |
| "learning_rate": 7.619929529850397e-05, | |
| "loss": 0.1926, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 1.2522639036178589, | |
| "eval_runtime": 9.6209, | |
| "eval_samples_per_second": 2.495, | |
| "eval_steps_per_second": 2.495, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 18.181818181818183, | |
| "grad_norm": 1.041843295097351, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.1788, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 18.545454545454547, | |
| "grad_norm": 1.2381653785705566, | |
| "learning_rate": 7.378121045351378e-05, | |
| "loss": 0.146, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "grad_norm": 1.218409776687622, | |
| "learning_rate": 7.254387703447154e-05, | |
| "loss": 0.1477, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "eval_loss": 1.470971941947937, | |
| "eval_runtime": 9.6218, | |
| "eval_samples_per_second": 2.494, | |
| "eval_steps_per_second": 2.494, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "step": 104, | |
| "total_flos": 2.5664889569673216e+16, | |
| "train_loss": 0.9187728297681763, | |
| "train_runtime": 1959.5505, | |
| "train_samples_per_second": 2.245, | |
| "train_steps_per_second": 0.128 | |
| } | |
| ], | |
| "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": 2.5664889569673216e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |