| { | |
| "best_metric": 0.11002114415168762, | |
| "best_model_checkpoint": "tuple-1k-t5/checkpoint-300", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 300, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 29.495084762573242, | |
| "learning_rate": 5e-06, | |
| "loss": 10.2171, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 37.105918884277344, | |
| "learning_rate": 1.1666666666666668e-05, | |
| "loss": 8.2109, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "grad_norm": 45.43325424194336, | |
| "learning_rate": 2e-05, | |
| "loss": 7.2074, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 4.679723739624023, | |
| "learning_rate": 2.8333333333333335e-05, | |
| "loss": 2.2548, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 3.3433218002319336, | |
| "learning_rate": 3.6666666666666666e-05, | |
| "loss": 1.315, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 3.405259370803833, | |
| "learning_rate": 4.5e-05, | |
| "loss": 1.0132, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "grad_norm": 2.41959810256958, | |
| "learning_rate": 4.962962962962963e-05, | |
| "loss": 0.7069, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 2.0366761684417725, | |
| "learning_rate": 4.8703703703703704e-05, | |
| "loss": 0.4676, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 1.5966047048568726, | |
| "learning_rate": 4.7777777777777784e-05, | |
| "loss": 0.325, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.965764045715332, | |
| "learning_rate": 4.685185185185185e-05, | |
| "loss": 0.2708, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 1.0757189989089966, | |
| "learning_rate": 4.592592592592593e-05, | |
| "loss": 0.241, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 0.9667235612869263, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.2218, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 0.9872345924377441, | |
| "learning_rate": 4.4074074074074076e-05, | |
| "loss": 0.205, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 0.7139382362365723, | |
| "learning_rate": 4.314814814814815e-05, | |
| "loss": 0.1784, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 0.9039368033409119, | |
| "learning_rate": 4.222222222222222e-05, | |
| "loss": 0.1794, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.6482570767402649, | |
| "learning_rate": 4.12962962962963e-05, | |
| "loss": 0.1837, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 0.7324709296226501, | |
| "learning_rate": 4.0370370370370374e-05, | |
| "loss": 0.1669, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 0.5565744638442993, | |
| "learning_rate": 3.944444444444445e-05, | |
| "loss": 0.164, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 0.5513789653778076, | |
| "learning_rate": 3.851851851851852e-05, | |
| "loss": 0.15, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.5113586187362671, | |
| "learning_rate": 3.759259259259259e-05, | |
| "loss": 0.149, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_gen_len": 19.0, | |
| "eval_loss": 0.13265687227249146, | |
| "eval_rouge1": 14.8515, | |
| "eval_rouge2": 12.024, | |
| "eval_rougeL": 14.8324, | |
| "eval_rougeLsum": 14.8405, | |
| "eval_runtime": 11.6075, | |
| "eval_samples_per_second": 17.23, | |
| "eval_steps_per_second": 1.12, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.05, | |
| "grad_norm": 0.6177830696105957, | |
| "learning_rate": 3.6666666666666666e-05, | |
| "loss": 0.1681, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 0.46891486644744873, | |
| "learning_rate": 3.574074074074074e-05, | |
| "loss": 0.1295, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.15, | |
| "grad_norm": 0.572552502155304, | |
| "learning_rate": 3.481481481481482e-05, | |
| "loss": 0.1357, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.4963364601135254, | |
| "learning_rate": 3.388888888888889e-05, | |
| "loss": 0.149, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "grad_norm": 0.538290798664093, | |
| "learning_rate": 3.2962962962962964e-05, | |
| "loss": 0.1358, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 0.528228223323822, | |
| "learning_rate": 3.203703703703704e-05, | |
| "loss": 0.1571, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "grad_norm": 0.7794693112373352, | |
| "learning_rate": 3.111111111111111e-05, | |
| "loss": 0.1361, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 0.5048784017562866, | |
| "learning_rate": 3.018518518518519e-05, | |
| "loss": 0.1376, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.45, | |
| "grad_norm": 0.568100094795227, | |
| "learning_rate": 2.925925925925926e-05, | |
| "loss": 0.1575, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 0.58453369140625, | |
| "learning_rate": 2.8333333333333335e-05, | |
| "loss": 0.1402, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "grad_norm": 0.4158877730369568, | |
| "learning_rate": 2.7407407407407408e-05, | |
| "loss": 0.1238, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 0.5241145491600037, | |
| "learning_rate": 2.6481481481481485e-05, | |
| "loss": 0.1318, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "grad_norm": 0.6674245595932007, | |
| "learning_rate": 2.5555555555555554e-05, | |
| "loss": 0.1347, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "grad_norm": 0.43407198786735535, | |
| "learning_rate": 2.462962962962963e-05, | |
| "loss": 0.1406, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "grad_norm": 0.5406764149665833, | |
| "learning_rate": 2.3703703703703707e-05, | |
| "loss": 0.1401, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 0.5779006481170654, | |
| "learning_rate": 2.277777777777778e-05, | |
| "loss": 0.1619, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 0.507857620716095, | |
| "learning_rate": 2.1851851851851852e-05, | |
| "loss": 0.1208, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 0.41192862391471863, | |
| "learning_rate": 2.0925925925925925e-05, | |
| "loss": 0.1303, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "grad_norm": 0.495255708694458, | |
| "learning_rate": 2e-05, | |
| "loss": 0.1224, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.3549606204032898, | |
| "learning_rate": 1.9074074074074075e-05, | |
| "loss": 0.1275, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_gen_len": 19.0, | |
| "eval_loss": 0.11091431230306625, | |
| "eval_rouge1": 14.8075, | |
| "eval_rouge2": 11.9873, | |
| "eval_rougeL": 14.7889, | |
| "eval_rougeLsum": 14.8017, | |
| "eval_runtime": 10.2766, | |
| "eval_samples_per_second": 19.462, | |
| "eval_steps_per_second": 1.265, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "grad_norm": 0.41942712664604187, | |
| "learning_rate": 1.814814814814815e-05, | |
| "loss": 0.1361, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.46660855412483215, | |
| "learning_rate": 1.7222222222222224e-05, | |
| "loss": 0.1144, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.15, | |
| "grad_norm": 0.5672628879547119, | |
| "learning_rate": 1.62962962962963e-05, | |
| "loss": 0.1363, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 0.4134034216403961, | |
| "learning_rate": 1.537037037037037e-05, | |
| "loss": 0.1309, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "grad_norm": 0.5884687304496765, | |
| "learning_rate": 1.4444444444444444e-05, | |
| "loss": 0.1329, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "grad_norm": 0.44717901945114136, | |
| "learning_rate": 1.3518518518518519e-05, | |
| "loss": 0.1219, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 0.6104301810264587, | |
| "learning_rate": 1.2592592592592592e-05, | |
| "loss": 0.1268, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 0.5338377952575684, | |
| "learning_rate": 1.1666666666666668e-05, | |
| "loss": 0.1154, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "grad_norm": 0.41602200269699097, | |
| "learning_rate": 1.074074074074074e-05, | |
| "loss": 0.1283, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.4216574728488922, | |
| "learning_rate": 9.814814814814815e-06, | |
| "loss": 0.1285, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "grad_norm": 0.4989544749259949, | |
| "learning_rate": 8.88888888888889e-06, | |
| "loss": 0.1246, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "grad_norm": 0.6025294065475464, | |
| "learning_rate": 7.962962962962963e-06, | |
| "loss": 0.1269, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "grad_norm": 0.38507816195487976, | |
| "learning_rate": 7.0370370370370375e-06, | |
| "loss": 0.1138, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "grad_norm": 0.3658128082752228, | |
| "learning_rate": 6.111111111111111e-06, | |
| "loss": 0.1237, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "grad_norm": 0.4835338592529297, | |
| "learning_rate": 5.185185185185185e-06, | |
| "loss": 0.1527, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 0.5483537316322327, | |
| "learning_rate": 4.2592592592592596e-06, | |
| "loss": 0.1222, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "grad_norm": 0.586328387260437, | |
| "learning_rate": 3.3333333333333333e-06, | |
| "loss": 0.1281, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.6712453961372375, | |
| "learning_rate": 2.4074074074074075e-06, | |
| "loss": 0.1384, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.95, | |
| "grad_norm": 0.37056586146354675, | |
| "learning_rate": 1.4814814814814817e-06, | |
| "loss": 0.1181, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.374318927526474, | |
| "learning_rate": 5.555555555555556e-07, | |
| "loss": 0.1206, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_gen_len": 19.0, | |
| "eval_loss": 0.11002114415168762, | |
| "eval_rouge1": 14.787, | |
| "eval_rouge2": 11.9709, | |
| "eval_rougeL": 14.7712, | |
| "eval_rougeLsum": 14.7825, | |
| "eval_runtime": 10.2272, | |
| "eval_samples_per_second": 19.556, | |
| "eval_steps_per_second": 1.271, | |
| "step": 300 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 300, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 641137509089280.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |