| { | |
| "best_metric": 0.06574746966362, | |
| "best_model_checkpoint": "./beans_outputs/checkpoint-520", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 650, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "grad_norm": 2.219902515411377, | |
| "learning_rate": 1.9692307692307696e-05, | |
| "loss": 1.0246, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 1.863097906112671, | |
| "learning_rate": 1.9384615384615386e-05, | |
| "loss": 0.9458, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 1.9998500347137451, | |
| "learning_rate": 1.907692307692308e-05, | |
| "loss": 0.8433, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 2.666048288345337, | |
| "learning_rate": 1.876923076923077e-05, | |
| "loss": 0.6995, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "grad_norm": 4.233841419219971, | |
| "learning_rate": 1.8461538461538465e-05, | |
| "loss": 0.663, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 2.192035436630249, | |
| "learning_rate": 1.8153846153846155e-05, | |
| "loss": 0.585, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 1.8312418460845947, | |
| "learning_rate": 1.784615384615385e-05, | |
| "loss": 0.5225, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "grad_norm": 1.2257707118988037, | |
| "learning_rate": 1.753846153846154e-05, | |
| "loss": 0.3896, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 2.517259359359741, | |
| "learning_rate": 1.7230769230769234e-05, | |
| "loss": 0.3501, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 3.061749219894409, | |
| "learning_rate": 1.6923076923076924e-05, | |
| "loss": 0.3671, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 1.758705973625183, | |
| "learning_rate": 1.6615384615384618e-05, | |
| "loss": 0.3307, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "grad_norm": 2.4563472270965576, | |
| "learning_rate": 1.630769230769231e-05, | |
| "loss": 0.2444, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 7.120365619659424, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 0.2814, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.9624060150375939, | |
| "eval_loss": 0.21468126773834229, | |
| "eval_runtime": 1.1414, | |
| "eval_samples_per_second": 116.529, | |
| "eval_steps_per_second": 14.895, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "grad_norm": 1.7768912315368652, | |
| "learning_rate": 1.5692307692307693e-05, | |
| "loss": 0.2, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 3.082838535308838, | |
| "learning_rate": 1.5384615384615387e-05, | |
| "loss": 0.2219, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "grad_norm": 1.386346459388733, | |
| "learning_rate": 1.5076923076923078e-05, | |
| "loss": 0.2001, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.31, | |
| "grad_norm": 0.5072073340415955, | |
| "learning_rate": 1.4769230769230772e-05, | |
| "loss": 0.2183, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "grad_norm": 0.6016378998756409, | |
| "learning_rate": 1.4461538461538462e-05, | |
| "loss": 0.2069, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "grad_norm": 2.8826181888580322, | |
| "learning_rate": 1.4153846153846156e-05, | |
| "loss": 0.1539, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "grad_norm": 0.43352940678596497, | |
| "learning_rate": 1.3846153846153847e-05, | |
| "loss": 0.2524, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "grad_norm": 0.5819545984268188, | |
| "learning_rate": 1.353846153846154e-05, | |
| "loss": 0.1836, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "grad_norm": 2.1941072940826416, | |
| "learning_rate": 1.3230769230769231e-05, | |
| "loss": 0.2198, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "grad_norm": 1.9567605257034302, | |
| "learning_rate": 1.2923076923076925e-05, | |
| "loss": 0.1557, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 2.7541744709014893, | |
| "learning_rate": 1.2615384615384616e-05, | |
| "loss": 0.2846, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "grad_norm": 0.38941043615341187, | |
| "learning_rate": 1.230769230769231e-05, | |
| "loss": 0.1364, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.34662145376205444, | |
| "learning_rate": 1.2e-05, | |
| "loss": 0.1326, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9699248120300752, | |
| "eval_loss": 0.12728814780712128, | |
| "eval_runtime": 1.0735, | |
| "eval_samples_per_second": 123.898, | |
| "eval_steps_per_second": 15.837, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "grad_norm": 0.4577018618583679, | |
| "learning_rate": 1.1692307692307694e-05, | |
| "loss": 0.0934, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 0.3331770598888397, | |
| "learning_rate": 1.1384615384615385e-05, | |
| "loss": 0.1416, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "grad_norm": 7.138263702392578, | |
| "learning_rate": 1.1076923076923079e-05, | |
| "loss": 0.1296, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "grad_norm": 8.594639778137207, | |
| "learning_rate": 1.076923076923077e-05, | |
| "loss": 0.1417, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "grad_norm": 5.681203365325928, | |
| "learning_rate": 1.0461538461538463e-05, | |
| "loss": 0.1707, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.46, | |
| "grad_norm": 0.3230597674846649, | |
| "learning_rate": 1.0153846153846154e-05, | |
| "loss": 0.1496, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.54, | |
| "grad_norm": 6.502106189727783, | |
| "learning_rate": 9.846153846153848e-06, | |
| "loss": 0.1758, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 1.060750126838684, | |
| "learning_rate": 9.53846153846154e-06, | |
| "loss": 0.0753, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "grad_norm": 0.3386625051498413, | |
| "learning_rate": 9.230769230769232e-06, | |
| "loss": 0.1345, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.77, | |
| "grad_norm": 6.454658031463623, | |
| "learning_rate": 8.923076923076925e-06, | |
| "loss": 0.117, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 0.3338567614555359, | |
| "learning_rate": 8.615384615384617e-06, | |
| "loss": 0.1511, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "grad_norm": 1.0864334106445312, | |
| "learning_rate": 8.307692307692309e-06, | |
| "loss": 0.1284, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.30264168977737427, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.1413, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.09794322401285172, | |
| "eval_runtime": 1.1225, | |
| "eval_samples_per_second": 118.486, | |
| "eval_steps_per_second": 15.145, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "grad_norm": 3.8275489807128906, | |
| "learning_rate": 7.692307692307694e-06, | |
| "loss": 0.1396, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.15, | |
| "grad_norm": 6.871574401855469, | |
| "learning_rate": 7.384615384615386e-06, | |
| "loss": 0.1577, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "grad_norm": 8.653078079223633, | |
| "learning_rate": 7.076923076923078e-06, | |
| "loss": 0.1429, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "grad_norm": 0.23040761053562164, | |
| "learning_rate": 6.76923076923077e-06, | |
| "loss": 0.0676, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "grad_norm": 4.652414798736572, | |
| "learning_rate": 6.461538461538463e-06, | |
| "loss": 0.1215, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.46, | |
| "grad_norm": 11.157665252685547, | |
| "learning_rate": 6.153846153846155e-06, | |
| "loss": 0.0796, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "grad_norm": 4.33862829208374, | |
| "learning_rate": 5.846153846153847e-06, | |
| "loss": 0.0967, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.62, | |
| "grad_norm": 7.9337921142578125, | |
| "learning_rate": 5.538461538461539e-06, | |
| "loss": 0.0716, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 3.69, | |
| "grad_norm": 0.22305496037006378, | |
| "learning_rate": 5.230769230769232e-06, | |
| "loss": 0.0902, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 3.77, | |
| "grad_norm": 0.2107217162847519, | |
| "learning_rate": 4.923076923076924e-06, | |
| "loss": 0.072, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 3.85, | |
| "grad_norm": 0.9404855370521545, | |
| "learning_rate": 4.615384615384616e-06, | |
| "loss": 0.1483, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "grad_norm": 0.21295365691184998, | |
| "learning_rate": 4.307692307692308e-06, | |
| "loss": 0.1349, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.3001614511013031, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.0855, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9849624060150376, | |
| "eval_loss": 0.06574746966362, | |
| "eval_runtime": 1.128, | |
| "eval_samples_per_second": 117.907, | |
| "eval_steps_per_second": 15.071, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "grad_norm": 8.45413589477539, | |
| "learning_rate": 3.692307692307693e-06, | |
| "loss": 0.0996, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 4.15, | |
| "grad_norm": 1.4845136404037476, | |
| "learning_rate": 3.384615384615385e-06, | |
| "loss": 0.1602, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "grad_norm": 0.35962048172950745, | |
| "learning_rate": 3.0769230769230774e-06, | |
| "loss": 0.1495, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.31, | |
| "grad_norm": 1.3418010473251343, | |
| "learning_rate": 2.7692307692307697e-06, | |
| "loss": 0.0511, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 4.38, | |
| "grad_norm": 0.33496251702308655, | |
| "learning_rate": 2.461538461538462e-06, | |
| "loss": 0.1355, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 4.46, | |
| "grad_norm": 0.2655811011791229, | |
| "learning_rate": 2.153846153846154e-06, | |
| "loss": 0.0836, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 4.54, | |
| "grad_norm": 10.997761726379395, | |
| "learning_rate": 1.8461538461538465e-06, | |
| "loss": 0.0673, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 4.62, | |
| "grad_norm": 2.0622777938842773, | |
| "learning_rate": 1.5384615384615387e-06, | |
| "loss": 0.1054, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "grad_norm": 5.778710842132568, | |
| "learning_rate": 1.230769230769231e-06, | |
| "loss": 0.092, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 4.77, | |
| "grad_norm": 0.2569364309310913, | |
| "learning_rate": 9.230769230769232e-07, | |
| "loss": 0.075, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 4.85, | |
| "grad_norm": 0.21514290571212769, | |
| "learning_rate": 6.153846153846155e-07, | |
| "loss": 0.0612, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "grad_norm": 0.23064370453357697, | |
| "learning_rate": 3.0769230769230774e-07, | |
| "loss": 0.0783, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.6160356402397156, | |
| "learning_rate": 0.0, | |
| "loss": 0.1234, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9774436090225563, | |
| "eval_loss": 0.08280191570520401, | |
| "eval_runtime": 1.0692, | |
| "eval_samples_per_second": 124.39, | |
| "eval_steps_per_second": 15.899, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 650, | |
| "total_flos": 4.006371770595533e+17, | |
| "train_loss": 0.2192855260005364, | |
| "train_runtime": 96.5552, | |
| "train_samples_per_second": 53.544, | |
| "train_steps_per_second": 6.732 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 650, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 4.006371770595533e+17, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |