| { | |
| "best_metric": 0.3667507469654083, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-3_120-samples_config-2_auto/checkpoint-44", | |
| "epoch": 14.909090909090908, | |
| "eval_steps": 500, | |
| "global_step": 82, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 8.562411308288574, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 3.7901, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 6.801224231719971, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 3.3615, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 6.11488676071167, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 3.1348, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 3.034968137741089, | |
| "eval_runtime": 7.6879, | |
| "eval_samples_per_second": 3.122, | |
| "eval_steps_per_second": 3.122, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 6.204690456390381, | |
| "learning_rate": 2.4e-05, | |
| "loss": 3.0712, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 7.684451103210449, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 2.9804, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 5.518642902374268, | |
| "learning_rate": 4e-05, | |
| "loss": 2.28, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.6218465566635132, | |
| "eval_runtime": 7.4994, | |
| "eval_samples_per_second": 3.2, | |
| "eval_steps_per_second": 3.2, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 3.9958155155181885, | |
| "learning_rate": 4.8e-05, | |
| "loss": 1.8747, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 3.0953969955444336, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.3616, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 2.8533308506011963, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.0138, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 0.8415324687957764, | |
| "eval_runtime": 7.4966, | |
| "eval_samples_per_second": 3.201, | |
| "eval_steps_per_second": 3.201, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 2.1823668479919434, | |
| "learning_rate": 7.2e-05, | |
| "loss": 0.6366, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.8621670603752136, | |
| "learning_rate": 8e-05, | |
| "loss": 0.6212, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.7150596976280212, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 0.6086, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.5121175646781921, | |
| "eval_runtime": 7.4909, | |
| "eval_samples_per_second": 3.204, | |
| "eval_steps_per_second": 3.204, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.8133587837219238, | |
| "learning_rate": 9.6e-05, | |
| "loss": 0.4973, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.48412278294563293, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 0.3577, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 0.43591466546058655, | |
| "eval_runtime": 7.4943, | |
| "eval_samples_per_second": 3.202, | |
| "eval_steps_per_second": 3.202, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.581091046333313, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 0.4331, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.40145689249038696, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.323, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.3245043158531189, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.3199, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.39703628420829773, | |
| "eval_runtime": 7.4999, | |
| "eval_samples_per_second": 3.2, | |
| "eval_steps_per_second": 3.2, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.29016050696372986, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.3172, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.4871235191822052, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.408, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.41161438822746277, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.2918, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.3742358386516571, | |
| "eval_runtime": 7.497, | |
| "eval_samples_per_second": 3.201, | |
| "eval_steps_per_second": 3.201, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.4374900460243225, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.3194, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.2966398596763611, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.1835, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.3199334442615509, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.2444, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.3667507469654083, | |
| "eval_runtime": 7.5133, | |
| "eval_samples_per_second": 3.194, | |
| "eval_steps_per_second": 3.194, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.35274624824523926, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.2349, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.389423131942749, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.2258, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.37230217456817627, | |
| "eval_runtime": 7.5018, | |
| "eval_samples_per_second": 3.199, | |
| "eval_steps_per_second": 3.199, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 1.3381692171096802, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.2493, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.7248063683509827, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.2156, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.412747859954834, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.1726, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.3866863250732422, | |
| "eval_runtime": 7.494, | |
| "eval_samples_per_second": 3.203, | |
| "eval_steps_per_second": 3.203, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.49216005206108093, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.1463, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.3493175208568573, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.1711, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.35457953810691833, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.1425, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.41673216223716736, | |
| "eval_runtime": 7.5017, | |
| "eval_samples_per_second": 3.199, | |
| "eval_steps_per_second": 3.199, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.4411362111568451, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.1458, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.3679543137550354, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.0949, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.7628819346427917, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.1228, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.4850577414035797, | |
| "eval_runtime": 7.4948, | |
| "eval_samples_per_second": 3.202, | |
| "eval_steps_per_second": 3.202, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.5807320475578308, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.0577, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.6347141265869141, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.103, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 0.5679486393928528, | |
| "eval_runtime": 7.4982, | |
| "eval_samples_per_second": 3.201, | |
| "eval_steps_per_second": 3.201, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.6395418643951416, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.0758, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.6982291340827942, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.0414, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.7121686935424805, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.0545, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.6688669323921204, | |
| "eval_runtime": 7.4887, | |
| "eval_samples_per_second": 3.205, | |
| "eval_steps_per_second": 3.205, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.38900575041770935, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.054, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 0.6852383017539978, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.038, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 0.3993234634399414, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.0273, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 0.729881763458252, | |
| "eval_runtime": 7.4911, | |
| "eval_samples_per_second": 3.204, | |
| "eval_steps_per_second": 3.204, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "step": 82, | |
| "total_flos": 3.5727052044763136e+16, | |
| "train_loss": 0.6642597818111138, | |
| "train_runtime": 1507.2702, | |
| "train_samples_per_second": 2.919, | |
| "train_steps_per_second": 0.166 | |
| } | |
| ], | |
| "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": 3.5727052044763136e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |