| { | |
| "best_metric": 0.0750546008348465, | |
| "best_model_checkpoint": "output/single-custom/quirky_sciq_raw/checkpoint-3000", | |
| "epoch": 2.3713071830846753, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 5.271481286241434e-07, | |
| "loss": 6.6565, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 1.0542962572482868e-06, | |
| "loss": 6.3909, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1.5814443858724301e-06, | |
| "loss": 5.9624, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 2.1085925144965737e-06, | |
| "loss": 4.7754, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 2.635740643120717e-06, | |
| "loss": 3.2357, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 3.1628887717448603e-06, | |
| "loss": 1.789, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 3.690036900369004e-06, | |
| "loss": 0.6753, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.217185028993147e-06, | |
| "loss": 0.382, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.744333157617291e-06, | |
| "loss": 0.3311, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 5.271481286241434e-06, | |
| "loss": 0.3008, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_loss": 0.29452574253082275, | |
| "eval_val_runtime": 79.8728, | |
| "eval_val_samples_per_second": 22.686, | |
| "eval_val_steps_per_second": 2.842, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_alice_loss": 0.16685596108436584, | |
| "eval_val_alice_runtime": 39.3274, | |
| "eval_val_alice_samples_per_second": 23.165, | |
| "eval_val_alice_steps_per_second": 2.899, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_bob_loss": 0.42446252703666687, | |
| "eval_val_bob_runtime": 38.7654, | |
| "eval_val_bob_samples_per_second": 23.242, | |
| "eval_val_bob_steps_per_second": 2.915, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_bob_gt_loss": 0.16960681974887848, | |
| "eval_val_bob_gt_runtime": 38.7527, | |
| "eval_val_bob_gt_samples_per_second": 23.25, | |
| "eval_val_bob_gt_steps_per_second": 2.916, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 5.798629414865578e-06, | |
| "loss": 0.2991, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 6.325777543489721e-06, | |
| "loss": 0.286, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 6.852925672113865e-06, | |
| "loss": 0.2509, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 7.380073800738008e-06, | |
| "loss": 0.2708, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 7.907221929362151e-06, | |
| "loss": 0.2296, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 8.434370057986295e-06, | |
| "loss": 0.2405, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 8.961518186610438e-06, | |
| "loss": 0.2393, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 9.488666315234582e-06, | |
| "loss": 0.2311, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 1.0015814443858725e-05, | |
| "loss": 0.2079, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 1.0542962572482869e-05, | |
| "loss": 0.1969, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_loss": 0.1757764369249344, | |
| "eval_val_runtime": 80.2328, | |
| "eval_val_samples_per_second": 22.584, | |
| "eval_val_steps_per_second": 2.829, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_alice_loss": 0.11201322823762894, | |
| "eval_val_alice_runtime": 39.4556, | |
| "eval_val_alice_samples_per_second": 23.089, | |
| "eval_val_alice_steps_per_second": 2.889, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_bob_loss": 0.2400250881910324, | |
| "eval_val_bob_runtime": 38.8802, | |
| "eval_val_bob_samples_per_second": 23.174, | |
| "eval_val_bob_steps_per_second": 2.906, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_bob_gt_loss": 0.15817159414291382, | |
| "eval_val_bob_gt_runtime": 38.8537, | |
| "eval_val_bob_gt_samples_per_second": 23.19, | |
| "eval_val_bob_gt_steps_per_second": 2.908, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 1.1070110701107012e-05, | |
| "loss": 0.1825, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 1.1597258829731156e-05, | |
| "loss": 0.1826, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.21244069583553e-05, | |
| "loss": 0.177, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 1.2651555086979441e-05, | |
| "loss": 0.1951, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 1.3178703215603585e-05, | |
| "loss": 0.2114, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 1.370585134422773e-05, | |
| "loss": 0.1543, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 1.4232999472851872e-05, | |
| "loss": 0.1454, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 1.4760147601476015e-05, | |
| "loss": 0.1458, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 1.528729573010016e-05, | |
| "loss": 0.1637, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 1.5814443858724302e-05, | |
| "loss": 0.1329, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_loss": 0.12391708046197891, | |
| "eval_val_runtime": 80.057, | |
| "eval_val_samples_per_second": 22.634, | |
| "eval_val_steps_per_second": 2.835, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_alice_loss": 0.11775108426809311, | |
| "eval_val_alice_runtime": 39.3993, | |
| "eval_val_alice_samples_per_second": 23.122, | |
| "eval_val_alice_steps_per_second": 2.893, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_bob_loss": 0.13003212213516235, | |
| "eval_val_bob_runtime": 38.8672, | |
| "eval_val_bob_samples_per_second": 23.182, | |
| "eval_val_bob_steps_per_second": 2.907, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_bob_gt_loss": 0.25759872794151306, | |
| "eval_val_bob_gt_runtime": 38.8511, | |
| "eval_val_bob_gt_samples_per_second": 23.191, | |
| "eval_val_bob_gt_steps_per_second": 2.909, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 1.6341591987348446e-05, | |
| "loss": 0.1552, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 1.686874011597259e-05, | |
| "loss": 0.1518, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 1.7395888244596733e-05, | |
| "loss": 0.1137, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 1.7923036373220876e-05, | |
| "loss": 0.1424, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 1.845018450184502e-05, | |
| "loss": 0.1131, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 1.8977332630469163e-05, | |
| "loss": 0.0887, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 1.9504480759093307e-05, | |
| "loss": 0.1046, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 1.9994420161815308e-05, | |
| "loss": 0.1071, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 1.99014228587371e-05, | |
| "loss": 0.092, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 1.980842555565889e-05, | |
| "loss": 0.1167, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_loss": 0.09470801800489426, | |
| "eval_val_runtime": 79.8646, | |
| "eval_val_samples_per_second": 22.688, | |
| "eval_val_steps_per_second": 2.842, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_alice_loss": 0.09601828455924988, | |
| "eval_val_alice_runtime": 39.3347, | |
| "eval_val_alice_samples_per_second": 23.16, | |
| "eval_val_alice_steps_per_second": 2.898, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_bob_loss": 0.09395376592874527, | |
| "eval_val_bob_runtime": 38.7824, | |
| "eval_val_bob_samples_per_second": 23.232, | |
| "eval_val_bob_steps_per_second": 2.914, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_bob_gt_loss": 0.4563288986682892, | |
| "eval_val_bob_gt_runtime": 38.7783, | |
| "eval_val_bob_gt_samples_per_second": 23.235, | |
| "eval_val_bob_gt_steps_per_second": 2.914, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 1.971542825258068e-05, | |
| "loss": 0.1076, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 1.9622430949502466e-05, | |
| "loss": 0.0904, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 1.9529433646424256e-05, | |
| "loss": 0.0718, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 1.9436436343346043e-05, | |
| "loss": 0.0889, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 1.9343439040267833e-05, | |
| "loss": 0.0987, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 1.9250441737189623e-05, | |
| "loss": 0.0782, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 1.9157444434111413e-05, | |
| "loss": 0.086, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 1.90644471310332e-05, | |
| "loss": 0.094, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "learning_rate": 1.897144982795499e-05, | |
| "loss": 0.0723, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 1.887845252487678e-05, | |
| "loss": 0.096, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_loss": 0.08209705352783203, | |
| "eval_val_runtime": 80.0462, | |
| "eval_val_samples_per_second": 22.637, | |
| "eval_val_steps_per_second": 2.836, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_alice_loss": 0.06850626319646835, | |
| "eval_val_alice_runtime": 39.4169, | |
| "eval_val_alice_samples_per_second": 23.112, | |
| "eval_val_alice_steps_per_second": 2.892, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_bob_loss": 0.09484687447547913, | |
| "eval_val_bob_runtime": 38.8822, | |
| "eval_val_bob_samples_per_second": 23.173, | |
| "eval_val_bob_steps_per_second": 2.906, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_bob_gt_loss": 0.43703609704971313, | |
| "eval_val_bob_gt_runtime": 38.8648, | |
| "eval_val_bob_gt_samples_per_second": 23.183, | |
| "eval_val_bob_gt_steps_per_second": 2.908, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 1.878545522179857e-05, | |
| "loss": 0.0841, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "learning_rate": 1.8692457918720358e-05, | |
| "loss": 0.0527, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 1.8599460615642148e-05, | |
| "loss": 0.0527, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1.8506463312563938e-05, | |
| "loss": 0.055, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1.8413466009485725e-05, | |
| "loss": 0.0642, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 1.8320468706407515e-05, | |
| "loss": 0.0574, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1.8227471403329305e-05, | |
| "loss": 0.0752, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 1.8134474100251092e-05, | |
| "loss": 0.0779, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.8041476797172882e-05, | |
| "loss": 0.0263, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "learning_rate": 1.7948479494094673e-05, | |
| "loss": 0.0461, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_val_loss": 0.0750546008348465, | |
| "eval_val_runtime": 80.0194, | |
| "eval_val_samples_per_second": 22.645, | |
| "eval_val_steps_per_second": 2.837, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_val_alice_loss": 0.07222797721624374, | |
| "eval_val_alice_runtime": 39.3985, | |
| "eval_val_alice_samples_per_second": 23.123, | |
| "eval_val_alice_steps_per_second": 2.894, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_val_bob_loss": 0.07870974391698837, | |
| "eval_val_bob_runtime": 38.8401, | |
| "eval_val_bob_samples_per_second": 23.198, | |
| "eval_val_bob_steps_per_second": 2.909, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.37, | |
| "eval_val_bob_gt_loss": 0.6858129501342773, | |
| "eval_val_bob_gt_runtime": 38.8232, | |
| "eval_val_bob_gt_samples_per_second": 23.208, | |
| "eval_val_bob_gt_steps_per_second": 2.911, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 12650, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.1701103468235817e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |