| { | |
| "best_metric": 0.08199693262577057, | |
| "best_model_checkpoint": "output/multi-custom/quirky_sciq_raw/checkpoint-2500", | |
| "epoch": 1.9760893192372295, | |
| "eval_steps": 500, | |
| "global_step": 2500, | |
| "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": 6.4515, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 1.0542962572482868e-06, | |
| "loss": 6.3183, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 1.5814443858724301e-06, | |
| "loss": 6.0849, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 2.1085925144965737e-06, | |
| "loss": 4.8196, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 2.635740643120717e-06, | |
| "loss": 3.3561, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 3.1628887717448603e-06, | |
| "loss": 1.7489, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 3.690036900369004e-06, | |
| "loss": 0.6441, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.217185028993147e-06, | |
| "loss": 0.3495, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.744333157617291e-06, | |
| "loss": 0.3412, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 5.271481286241434e-06, | |
| "loss": 0.3092, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_loss": 0.2828400433063507, | |
| "eval_val_runtime": 80.678, | |
| "eval_val_samples_per_second": 22.46, | |
| "eval_val_steps_per_second": 2.814, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_alice_loss": 0.1860445737838745, | |
| "eval_val_alice_runtime": 39.862, | |
| "eval_val_alice_samples_per_second": 22.854, | |
| "eval_val_alice_steps_per_second": 2.86, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_bob_loss": 0.38034239411354065, | |
| "eval_val_bob_runtime": 39.1823, | |
| "eval_val_bob_samples_per_second": 22.995, | |
| "eval_val_bob_steps_per_second": 2.884, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_val_bob_gt_loss": 0.17556993663311005, | |
| "eval_val_bob_gt_runtime": 39.1726, | |
| "eval_val_bob_gt_samples_per_second": 23.001, | |
| "eval_val_bob_gt_steps_per_second": 2.885, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 5.798629414865578e-06, | |
| "loss": 0.3015, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 6.325777543489721e-06, | |
| "loss": 0.2848, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 6.852925672113865e-06, | |
| "loss": 0.2623, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 7.380073800738008e-06, | |
| "loss": 0.2521, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 7.907221929362151e-06, | |
| "loss": 0.2507, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 8.434370057986295e-06, | |
| "loss": 0.2376, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 8.961518186610438e-06, | |
| "loss": 0.2749, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 9.488666315234582e-06, | |
| "loss": 0.2565, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 1.0015814443858725e-05, | |
| "loss": 0.2341, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 1.0542962572482869e-05, | |
| "loss": 0.2243, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_loss": 0.2023884356021881, | |
| "eval_val_runtime": 80.6373, | |
| "eval_val_samples_per_second": 22.471, | |
| "eval_val_steps_per_second": 2.815, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_alice_loss": 0.12862880527973175, | |
| "eval_val_alice_runtime": 39.8479, | |
| "eval_val_alice_samples_per_second": 22.862, | |
| "eval_val_alice_steps_per_second": 2.861, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_bob_loss": 0.2769617438316345, | |
| "eval_val_bob_runtime": 39.167, | |
| "eval_val_bob_samples_per_second": 23.004, | |
| "eval_val_bob_steps_per_second": 2.885, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_val_bob_gt_loss": 0.1369783878326416, | |
| "eval_val_bob_gt_runtime": 39.147, | |
| "eval_val_bob_gt_samples_per_second": 23.016, | |
| "eval_val_bob_gt_steps_per_second": 2.887, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 1.1070110701107012e-05, | |
| "loss": 0.1934, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 1.1597258829731156e-05, | |
| "loss": 0.2017, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.21244069583553e-05, | |
| "loss": 0.1957, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 1.2651555086979441e-05, | |
| "loss": 0.2234, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 1.3178703215603585e-05, | |
| "loss": 0.2106, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 1.370585134422773e-05, | |
| "loss": 0.1806, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 1.4232999472851872e-05, | |
| "loss": 0.1805, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 1.4760147601476015e-05, | |
| "loss": 0.168, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "learning_rate": 1.528729573010016e-05, | |
| "loss": 0.185, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 1.5814443858724302e-05, | |
| "loss": 0.1657, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_loss": 0.15653647482395172, | |
| "eval_val_runtime": 80.515, | |
| "eval_val_samples_per_second": 22.505, | |
| "eval_val_steps_per_second": 2.819, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_alice_loss": 0.14135360717773438, | |
| "eval_val_alice_runtime": 39.8181, | |
| "eval_val_alice_samples_per_second": 22.879, | |
| "eval_val_alice_steps_per_second": 2.863, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_bob_loss": 0.17198161780834198, | |
| "eval_val_bob_runtime": 39.1191, | |
| "eval_val_bob_samples_per_second": 23.032, | |
| "eval_val_bob_steps_per_second": 2.889, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "eval_val_bob_gt_loss": 0.15408484637737274, | |
| "eval_val_bob_gt_runtime": 39.1152, | |
| "eval_val_bob_gt_samples_per_second": 23.035, | |
| "eval_val_bob_gt_steps_per_second": 2.889, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.23, | |
| "learning_rate": 1.6341591987348446e-05, | |
| "loss": 0.1991, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.26, | |
| "learning_rate": 1.686874011597259e-05, | |
| "loss": 0.1677, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 1.7395888244596733e-05, | |
| "loss": 0.1423, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 1.7923036373220876e-05, | |
| "loss": 0.1564, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 1.845018450184502e-05, | |
| "loss": 0.1324, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 1.8977332630469163e-05, | |
| "loss": 0.1181, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.46, | |
| "learning_rate": 1.9504480759093307e-05, | |
| "loss": 0.1293, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 1.9994420161815308e-05, | |
| "loss": 0.1327, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 1.99014228587371e-05, | |
| "loss": 0.1066, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 1.980842555565889e-05, | |
| "loss": 0.1511, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_loss": 0.10322380810976028, | |
| "eval_val_runtime": 80.72, | |
| "eval_val_samples_per_second": 22.448, | |
| "eval_val_steps_per_second": 2.812, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_alice_loss": 0.08724239468574524, | |
| "eval_val_alice_runtime": 39.8825, | |
| "eval_val_alice_samples_per_second": 22.842, | |
| "eval_val_alice_steps_per_second": 2.858, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_bob_loss": 0.11934794485569, | |
| "eval_val_bob_runtime": 39.2159, | |
| "eval_val_bob_samples_per_second": 22.975, | |
| "eval_val_bob_steps_per_second": 2.881, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "eval_val_bob_gt_loss": 0.3382713794708252, | |
| "eval_val_bob_gt_runtime": 39.1957, | |
| "eval_val_bob_gt_samples_per_second": 22.987, | |
| "eval_val_bob_gt_steps_per_second": 2.883, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 1.971542825258068e-05, | |
| "loss": 0.1052, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 1.9622430949502466e-05, | |
| "loss": 0.1176, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 1.9529433646424256e-05, | |
| "loss": 0.0829, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 1.74, | |
| "learning_rate": 1.9436436343346043e-05, | |
| "loss": 0.1116, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 1.9343439040267833e-05, | |
| "loss": 0.1181, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 1.9250441737189623e-05, | |
| "loss": 0.0796, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 1.9157444434111413e-05, | |
| "loss": 0.1005, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "learning_rate": 1.90644471310332e-05, | |
| "loss": 0.1047, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.94, | |
| "learning_rate": 1.897144982795499e-05, | |
| "loss": 0.0862, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 1.887845252487678e-05, | |
| "loss": 0.1224, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_loss": 0.08199693262577057, | |
| "eval_val_runtime": 80.6377, | |
| "eval_val_samples_per_second": 22.471, | |
| "eval_val_steps_per_second": 2.815, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_alice_loss": 0.059702351689338684, | |
| "eval_val_alice_runtime": 39.8524, | |
| "eval_val_alice_samples_per_second": 22.859, | |
| "eval_val_alice_steps_per_second": 2.861, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_bob_loss": 0.10402241349220276, | |
| "eval_val_bob_runtime": 39.1578, | |
| "eval_val_bob_samples_per_second": 23.009, | |
| "eval_val_bob_steps_per_second": 2.886, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "eval_val_bob_gt_loss": 0.4154125154018402, | |
| "eval_val_bob_gt_runtime": 39.1459, | |
| "eval_val_bob_gt_samples_per_second": 23.016, | |
| "eval_val_bob_gt_steps_per_second": 2.887, | |
| "step": 2500 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 12650, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 9.792119282851676e+17, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |