| { | |
| "best_metric": 1.8157883882522583, | |
| "best_model_checkpoint": "/tmp/tst-gun-gub-pt/checkpoint-40000", | |
| "epoch": 3.0, | |
| "eval_steps": 4000, | |
| "global_step": 40752, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 4.938653317628583e-05, | |
| "loss": 3.3185, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 4.877306635257165e-05, | |
| "loss": 2.967, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.815959952885748e-05, | |
| "loss": 2.803, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.754613270514331e-05, | |
| "loss": 2.6964, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.6932665881429136e-05, | |
| "loss": 2.6199, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.631919905771496e-05, | |
| "loss": 2.5435, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.5705732234000786e-05, | |
| "loss": 2.4687, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.509226541028662e-05, | |
| "loss": 2.4421, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "eval_bleu": 3.3785, | |
| "eval_gen_len": 68.48, | |
| "eval_loss": 2.5284605026245117, | |
| "eval_runtime": 39.0713, | |
| "eval_samples_per_second": 6.399, | |
| "eval_steps_per_second": 0.41, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.4478798586572437e-05, | |
| "loss": 2.3817, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 4.386533176285827e-05, | |
| "loss": 2.3424, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 4.3251864939144094e-05, | |
| "loss": 2.3047, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 4.263839811542992e-05, | |
| "loss": 2.2823, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 4.2024931291715744e-05, | |
| "loss": 2.2436, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.141146446800157e-05, | |
| "loss": 2.2266, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 4.07979976442874e-05, | |
| "loss": 2.1898, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 4.018453082057323e-05, | |
| "loss": 2.1667, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_bleu": 4.5883, | |
| "eval_gen_len": 58.6, | |
| "eval_loss": 2.3017916679382324, | |
| "eval_runtime": 34.6446, | |
| "eval_samples_per_second": 7.216, | |
| "eval_steps_per_second": 0.462, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 3.957106399685905e-05, | |
| "loss": 2.1403, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 3.895759717314488e-05, | |
| "loss": 2.1188, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 3.834413034943071e-05, | |
| "loss": 2.1002, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 3.773066352571653e-05, | |
| "loss": 2.0948, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 3.711719670200235e-05, | |
| "loss": 2.0678, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 3.6503729878288185e-05, | |
| "loss": 2.0502, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.589026305457401e-05, | |
| "loss": 2.0416, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 3.5276796230859836e-05, | |
| "loss": 2.0255, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "eval_bleu": 5.1052, | |
| "eval_gen_len": 67.3, | |
| "eval_loss": 2.1290316581726074, | |
| "eval_runtime": 40.7019, | |
| "eval_samples_per_second": 6.142, | |
| "eval_steps_per_second": 0.393, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 3.466332940714566e-05, | |
| "loss": 2.0057, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 3.404986258343149e-05, | |
| "loss": 1.9979, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 3.343639575971731e-05, | |
| "loss": 1.9726, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 1.03, | |
| "learning_rate": 3.2822928936003144e-05, | |
| "loss": 1.9261, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 3.220946211228897e-05, | |
| "loss": 1.9133, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 3.15959952885748e-05, | |
| "loss": 1.9109, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 3.098252846486062e-05, | |
| "loss": 1.9171, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 3.0369061641146445e-05, | |
| "loss": 1.8995, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "eval_bleu": 7.8429, | |
| "eval_gen_len": 55.48, | |
| "eval_loss": 2.0535314083099365, | |
| "eval_runtime": 32.581, | |
| "eval_samples_per_second": 7.673, | |
| "eval_steps_per_second": 0.491, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 2.9755594817432277e-05, | |
| "loss": 1.8771, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 2.91421279937181e-05, | |
| "loss": 1.8841, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 2.8528661170003927e-05, | |
| "loss": 1.8787, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 2.7915194346289753e-05, | |
| "loss": 1.8647, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 2.730172752257558e-05, | |
| "loss": 1.8488, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 2.6688260698861407e-05, | |
| "loss": 1.8342, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.44, | |
| "learning_rate": 2.6074793875147235e-05, | |
| "loss": 1.833, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 2.546132705143306e-05, | |
| "loss": 1.8322, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "eval_bleu": 7.2663, | |
| "eval_gen_len": 58.24, | |
| "eval_loss": 1.9960261583328247, | |
| "eval_runtime": 35.7526, | |
| "eval_samples_per_second": 6.993, | |
| "eval_steps_per_second": 0.448, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.51, | |
| "learning_rate": 2.4847860227718886e-05, | |
| "loss": 1.8187, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 1.55, | |
| "learning_rate": 2.4234393404004714e-05, | |
| "loss": 1.8118, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 2.362092658029054e-05, | |
| "loss": 1.8144, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 1.62, | |
| "learning_rate": 2.3007459756576365e-05, | |
| "loss": 1.7989, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "learning_rate": 2.2393992932862194e-05, | |
| "loss": 1.8068, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 1.69, | |
| "learning_rate": 2.178052610914802e-05, | |
| "loss": 1.8043, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 1.73, | |
| "learning_rate": 2.1167059285433844e-05, | |
| "loss": 1.7863, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 2.055359246171967e-05, | |
| "loss": 1.7868, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "eval_bleu": 7.0981, | |
| "eval_gen_len": 66.34, | |
| "eval_loss": 1.922366976737976, | |
| "eval_runtime": 40.7924, | |
| "eval_samples_per_second": 6.129, | |
| "eval_steps_per_second": 0.392, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "learning_rate": 1.9940125638005498e-05, | |
| "loss": 1.7759, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 1.9326658814291323e-05, | |
| "loss": 1.7653, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "learning_rate": 1.871319199057715e-05, | |
| "loss": 1.7689, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 1.8099725166862977e-05, | |
| "loss": 1.758, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 1.95, | |
| "learning_rate": 1.7486258343148802e-05, | |
| "loss": 1.754, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 1.687279151943463e-05, | |
| "loss": 1.7393, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 1.6259324695720456e-05, | |
| "loss": 1.7151, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "learning_rate": 1.5645857872006285e-05, | |
| "loss": 1.7012, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.06, | |
| "eval_bleu": 7.5657, | |
| "eval_gen_len": 60.3, | |
| "eval_loss": 1.8868523836135864, | |
| "eval_runtime": 38.0123, | |
| "eval_samples_per_second": 6.577, | |
| "eval_steps_per_second": 0.421, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "learning_rate": 1.503239104829211e-05, | |
| "loss": 1.6993, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 1.4418924224577934e-05, | |
| "loss": 1.6944, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1.380545740086376e-05, | |
| "loss": 1.7064, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 2.21, | |
| "learning_rate": 1.3191990577149588e-05, | |
| "loss": 1.6993, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1.2578523753435415e-05, | |
| "loss": 1.6893, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 1.1965056929721242e-05, | |
| "loss": 1.6874, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 2.32, | |
| "learning_rate": 1.1351590106007069e-05, | |
| "loss": 1.6793, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 1.0738123282292894e-05, | |
| "loss": 1.6773, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "eval_bleu": 7.9888, | |
| "eval_gen_len": 61.18, | |
| "eval_loss": 1.8613367080688477, | |
| "eval_runtime": 38.7116, | |
| "eval_samples_per_second": 6.458, | |
| "eval_steps_per_second": 0.413, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 1.012465645857872e-05, | |
| "loss": 1.6798, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 9.511189634864546e-06, | |
| "loss": 1.6823, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 2.47, | |
| "learning_rate": 8.897722811150373e-06, | |
| "loss": 1.6711, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 8.2842559874362e-06, | |
| "loss": 1.679, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.54, | |
| "learning_rate": 7.670789163722027e-06, | |
| "loss": 1.6662, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "learning_rate": 7.057322340007853e-06, | |
| "loss": 1.6716, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.61, | |
| "learning_rate": 6.4438555162936784e-06, | |
| "loss": 1.6569, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "learning_rate": 5.830388692579505e-06, | |
| "loss": 1.6631, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.65, | |
| "eval_bleu": 8.0862, | |
| "eval_gen_len": 60.5, | |
| "eval_loss": 1.8354450464248657, | |
| "eval_runtime": 38.0466, | |
| "eval_samples_per_second": 6.571, | |
| "eval_steps_per_second": 0.421, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "learning_rate": 5.216921868865332e-06, | |
| "loss": 1.6602, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 4.6034550451511585e-06, | |
| "loss": 1.6549, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "learning_rate": 3.989988221436985e-06, | |
| "loss": 1.6521, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "learning_rate": 3.3765213977228115e-06, | |
| "loss": 1.6643, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 2.7630545740086376e-06, | |
| "loss": 1.6533, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 2.87, | |
| "learning_rate": 2.149587750294464e-06, | |
| "loss": 1.6596, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 2.91, | |
| "learning_rate": 1.5361209265802905e-06, | |
| "loss": 1.6664, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 9.22654102866117e-07, | |
| "loss": 1.6379, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "eval_bleu": 8.4077, | |
| "eval_gen_len": 60.18, | |
| "eval_loss": 1.8157883882522583, | |
| "eval_runtime": 35.7813, | |
| "eval_samples_per_second": 6.987, | |
| "eval_steps_per_second": 0.447, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "learning_rate": 3.091872791519435e-07, | |
| "loss": 1.6493, | |
| "step": 40500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 40752, | |
| "total_flos": 3.4728164062396416e+16, | |
| "train_loss": 1.939443588256836, | |
| "train_runtime": 15380.8623, | |
| "train_samples_per_second": 42.392, | |
| "train_steps_per_second": 2.65 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 40752, | |
| "num_train_epochs": 3, | |
| "save_steps": 4000, | |
| "total_flos": 3.4728164062396416e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |